fix: unique str

Co-authored-by: minwoopark <minwoopark@google.com>

.cloud-build/cleanup/cleanup.py

@@ -7,6 +7,13 @@ from resource_cleanup_manager import (
     ResourceCleanupManager,
     MatchingEngineIndexEndpointResourceCleanupManager,
     MatchingEngineIndexResourceCleanupManager,
+    FeatureStoreCleanupManager,
+    PipelineJobCleanupManager,
+    TrainingJobCleanupManager,
+    HyperparameterTuningCleanupManager,
+    BatchPredictionJobCleanupManager,
+    ExperimentCleanupManager,
+    BucketCleanupManager
 )
 
 rate_limit = RateLimit(max_count=25, per=60, greedy=False)
@@ -23,7 +30,6 @@ def run_cleanup_managers(managers: List[ResourceCleanupManager], is_dry_run: boo
             try:
                 if not manager.is_deletable(resource):
                     continue
-
                 if is_dry_run:
                     resource_name = manager.resource_name(resource)
                     print(f"Will delete '{type_name}': {resource_name}")
@@ -48,6 +54,13 @@ managers: List[ResourceCleanupManager] = [
     ModelResourceCleanupManager(),  # ModelResourceCleanupManager must follow EndpointResourceCleanupManager due to deployed models blocking model deletion.
     MatchingEngineIndexEndpointResourceCleanupManager(),
     MatchingEngineIndexResourceCleanupManager(),
+    FeatureStoreCleanupManager(),
+    PipelineJobCleanupManager(),
+    TrainingJobCleanupManager(),
+    HyperparameterTuningCleanupManager(),
+    BatchPredictionJobCleanupManager(),
+    # ExperimentCleanupManager(), # Experiment missing _resource_noun
+    BucketCleanupManager()
 ]
 
 run_cleanup_managers(managers=managers, is_dry_run=is_dry_run)

.cloud-build/cleanup/resource_cleanup_manager.py

@@ -1,8 +1,17 @@
+'''
+READ FIRST BEFORE MAKING CHANGES
+-  Create a convention for resources created from vertex-ai-samples GH. We already have one IIRC
+- Only delete those objects as part of our clean-up script.
+- Don't run any tests on python-docs-samples-tests project, especially ones that affect resources created outside of our purview
+- Add --dry-run option to the clean-up script. This option will just output the list of resources the script will delete instead of actually deleting the resources.
+- Have a larger conversation in DEE before touching any resources that were not created as part of vertex-ai-samples
+'''
 import abc
 from typing import Any, Type
 
 from google.cloud import aiplatform
 from google.cloud.aiplatform import base
+from google.cloud import storage
 from proto.datetime_helpers import DatetimeWithNanoseconds
 
 # If a resource was updated within this number of seconds, do not delete.
@@ -69,7 +78,7 @@ class VertexAIResourceCleanupManager(ResourceCleanupManager):
     def delete(self, resource):
         resource.delete()
 
-    def get_seconds_since_modification(self, resource: Any) -> bool:
+    def get_seconds_since_modification(self, resource: Any) -> float:
         update_time = resource.update_time
         current_time = DatetimeWithNanoseconds.now(tz=update_time.tzinfo)
         return (current_time - update_time).total_seconds()
@@ -97,13 +106,10 @@ class EndpointResourceCleanupManager(VertexAIResourceCleanupManager):
     vertex_ai_resource = aiplatform.Endpoint
 
     def delete(self, resource):
-        # TODO: Remove this once https://github.com/googleapis/python-aiplatform/issues/1441 is fixed
-        resource._sync_gca_resource()
         for deployed_model_id in [
             models.id for models in resource._gca_resource.deployed_models
         ]:
             resource._undeploy(deployed_model_id=deployed_model_id)
-
         resource.delete(force=True)
 
 
@@ -117,3 +123,87 @@ class MatchingEngineIndexResourceCleanupManager(VertexAIResourceCleanupManager):
 
 class MatchingEngineIndexEndpointResourceCleanupManager(VertexAIResourceCleanupManager):
     vertex_ai_resource = aiplatform.MatchingEngineIndexEndpoint
+
+    def delete(self, resource):
+        resource.undeploy_all()
+        resource.delete(force=True)
+
+class FeatureStoreCleanupManager(VertexAIResourceCleanupManager):
+    vertex_ai_resource = aiplatform.Featurestore
+
+class PipelineJobCleanupManager(VertexAIResourceCleanupManager):
+    vertex_ai_resource = aiplatform.PipelineJob
+
+class TrainingJobCleanupManager(VertexAIResourceCleanupManager):
+    vertex_ai_resource = aiplatform.training_jobs._CustomTrainingJob
+
+    job_types = [
+            aiplatform.AutoMLImageTrainingJob,
+            aiplatform.AutoMLTextTrainingJob,
+            aiplatform.AutoMLTabularTrainingJob,
+            aiplatform.AutoMLVideoTrainingJob,
+            aiplatform.AutoMLForecastingTrainingJob,
+            aiplatform.CustomJob,
+            aiplatform.CustomTrainingJob,
+            aiplatform.CustomContainerTrainingJob,
+            aiplatform.CustomPythonPackageTrainingJob
+    ]
+
+    def list(self) -> Any:
+        return [
+            job
+            for job_type in self.job_types
+            for job in job_type.list()
+        ]
+
+class HyperparameterTuningCleanupManager(VertexAIResourceCleanupManager):
+    vertex_ai_resource = aiplatform.HyperparameterTuningJob
+
+
+class BatchPredictionJobCleanupManager(VertexAIResourceCleanupManager):
+    vertex_ai_resource = aiplatform.BatchPredictionJob
+
+class ExperimentCleanupManager(VertexAIResourceCleanupManager):
+    vertex_ai_resource = aiplatform.Experiment
+
+class BucketCleanupManager(ResourceCleanupManager):
+    vertex_ai_resource = storage.bucket.Bucket
+
+    def list(self) -> Any:
+        storage_client = storage.Client()
+        return list(storage_client.list_buckets())
+
+    def delete(self, resource):
+        try:
+            resource.delete(force=True)
+        except Exception as e:
+            print(e)
+
+    @property
+    def type_name(self) -> str:
+        return str(type(self.vertex_ai_resource))
+
+    def get_seconds_since_modification(self, resource: Any) -> float:
+        # Bucket has no last_update property, only time created
+        created_time = resource.time_created
+        current_time = DatetimeWithNanoseconds.now()
+        return float(current_time.timestamp() - created_time.timestamp())
+
+    def resource_name(self, resource: Any) -> str:
+        return resource.name
+
+    def is_deletable(self, resource: Any) -> bool:
+        time_difference = self.get_seconds_since_modification(resource)
+
+        if self.resource_name(resource).startswith('your-bucket-name'):
+            print(f"Skipping '{resource}' not a Vertex AI notebook bucket")
+            return False
+
+        # Check that it wasn't created too recently, to prevent race conditions
+        if time_difference <= RESOURCE_UPDATE_BUFFER_IN_SECONDS:
+            print(
+                f"Skipping '{resource}' due to update_time being '{time_difference}', which is less than '{RESOURCE_UPDATE_BUFFER_IN_SECONDS}'."
+            )
+            return False
+
+        return True

.cloud-build/execute_changed_notebooks_helper.py

@@ -156,7 +156,7 @@ def _create_tag(filepath: str) -> str:
     return tag
 
 
-rate_limit = RateLimit(max_count=50, per=60, greedy=True)
+rate_limit = RateLimit(max_count=10, per=60, greedy=True)
 
 
 def process_and_execute_notebook(
@@ -245,7 +245,7 @@ def process_and_execute_notebook(
         result.logs_bucket = operation_metadata.build.logs_bucket
 
         # Block and wait for the result
-        operation_result = operation.result()
+        operation_result = operation.result(timeout=timeout_in_seconds)
 
         result.duration = datetime.datetime.now() - time_start
         result.is_pass = True

.cloud-build/requirements.txt

@@ -10,4 +10,5 @@ google-cloud-aiplatform
 google-cloud-storage
 google-cloud-build
 ratemate
-GitPython
+GitPython
+tqdm

.github/workflows/linter/requirements.txt

@@ -2,9 +2,9 @@ git+https://github.com/tensorflow/docs
 ipython
 jupyter
 nbconvert
-black==22.10.0
+black==22.12.0
 pyupgrade==2.38.4
-isort==5.10.1
-flake8==4.0.1
+isort==5.12.0
+flake8==6.0.0
 nbqa==1.5.3
 

CONTRIBUTING.md

@@ -44,12 +44,10 @@ Finally, run this code block to check for errors. Each step will attempt to
 automatically fix any issues. If the fixes can't be performed automatically,
 then you will need to manually address them before submitting your PR.
 
+Note: For official, only submit one notebook per PR.
+
 ```shell
-nbqa black "$notebook"
-nbqa pyupgrade "$notebook"
-nbqa isort "$notebook"
-nbqa flake8 "$notebook" --extend-ignore=W391,E501,F821,E402,F404,W503,E203,E722,W293,W291
-python3 -m tensorflow_docs.tools.nbfmt --remove_outputs "$notebook"
+docker run -v ${PWD}:/setup/app gcr.io/cloud-devrel-public-resources/notebook_linter:latest your_notebook
 ```
 
 ## Code Reviews

community-content/CODEOWNERS

@@ -8,3 +8,5 @@
 /cpr-examples @samthrasher
 /Train_tabular_models_with_many_frameworks_and_import_to_Vertex_AI_using_Pipelines @Ark-kun
 /pipeline_components @Ark-kun
+/pipeline_components/image_ml_model_training @lakeyk
+/prediction_featurestore_integration @googleapis/vertex-prediction-team

community-content/Train_tabular_models_with_many_frameworks_and_import_to_Vertex_AI_using_Pipelines/Train_tabular_classification_logistic_regression_model_using_Scikit_learn_and_import_to_Vertex_AI/pipeline.py

@@ -6,8 +6,8 @@ download_from_gcs_op = components.load_component_from_url("https://raw.githubuse
 select_columns_using_Pandas_on_CSV_data_op = components.load_component_from_url("https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/399405402d95f4a011e2d2e967c96f8508ba5688/community-content/pipeline_components/pandas/Select_columns/in_CSV_format/component.yaml")
 fill_all_missing_values_using_Pandas_on_CSV_data_op = components.load_component_from_url("https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/399405402d95f4a011e2d2e967c96f8508ba5688/community-content/pipeline_components/pandas/Fill_all_missing_values/in_CSV_format/component.yaml")
 binarize_column_using_Pandas_on_CSV_data_op = components.load_component_from_url("https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/399405402d95f4a011e2d2e967c96f8508ba5688/community-content/pipeline_components/pandas/Binarize_column/in_CSV_format/component.yaml")
-train_logistic_regression_model_using_scikit_learn_from_CSV_op = components.load_component_from_url("https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/399405402d95f4a011e2d2e967c96f8508ba5688/community-content/pipeline_components/ML_frameworks/Scikit_learn/Train_logistic_regression_model/from_CSV/component.yaml")
-upload_Scikit_learn_pickle_model_to_Google_Cloud_Vertex_AI_op = components.load_component_from_url("https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/1f5cf6e06409b704064b2086c0a705e4e6b4fcde/community-content/pipeline_components/google-cloud/Vertex_AI/Models/Upload_Scikit-learn_pickle_model/component.yaml")
+train_logistic_regression_model_using_scikit_learn_from_CSV_op = components.load_component_from_url("https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/1f5cf6e06409b704064b2086c0a705e4e6b4fcde/community-content/pipeline_components/ML_frameworks/Scikit_learn/Train_logistic_regression_model/from_CSV/component.yaml")
+upload_Scikit_learn_pickle_model_to_Google_Cloud_Vertex_AI_op = components.load_component_from_url("https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/399405402d95f4a011e2d2e967c96f8508ba5688/community-content/pipeline_components/google-cloud/Vertex_AI/Models/Upload_Scikit-learn_pickle_model/component.yaml")
 deploy_model_to_endpoint_op = components.load_component_from_url("https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/399405402d95f4a011e2d2e967c96f8508ba5688/community-content/pipeline_components/google-cloud/Vertex_AI/Models/Deploy_to_endpoint/component.yaml")
 
 # %% Pipeline definition

community-content/pipeline_components/image_ml_model_training/load_image_classification_model/component.yaml

@@ -0,0 +1,112 @@
+name: Load image classification model from tfhub
+description: |
+  Loads specified model from TFHub, creates layer to receive additional (3 channel) imagery data.
+  Args:
+    class_names (Sequence[str]):
+        Sequence of strings of categories for classification corresponding to input data.
+    loaded_model_path (str):
+        Output path for the loaded model.
+    image_size_path (str):
+        Output path for the model expected image size.
+    model_name (Optional[str]):
+        Name of the pre-trained image classification model to load from TFHub.
+        Eligible model_name:
+        - efficientnetv2-s
+        - efficientnetv2-m
+        - efficientnetv2-l
+        - efficientnetv2-s-21k
+        - efficientnetv2-m-21k
+        - efficientnetv2-l-21k
+        - efficientnetv2-xl-21k
+        - efficientnetv2-b0-21k
+        - efficientnetv2-b1-21k
+        - efficientnetv2-b2-21k
+        - efficientnetv2-b3-21k
+        - efficientnetv2-s-21k-ft1k
+        - efficientnetv2-m-21k-ft1k
+        - efficientnetv2-l-21k-ft1k
+        - efficientnetv2-xl-21k-ft1k
+        - efficientnetv2-b0-21k-ft1k
+        - efficientnetv2-b1-21k-ft1k
+        - efficientnetv2-b2-21k-ft1k
+        - efficientnetv2-b3-21k-ft1k
+        - efficientnetv2-b0
+        - efficientnetv2-b1
+        - efficientnetv2-b2
+        - efficientnetv2-b3
+        - efficientnet_b0
+        - efficientnet_b1
+        - efficientnet_b2
+        - efficientnet_b3
+        - efficientnet_b4
+        - efficientnet_b5
+        - efficientnet_b6
+        - efficientnet_b7
+        - bit_s-r50x1
+        - inception_v3
+        - inception_resnet_v2
+        - resnet_v1_50
+        - resnet_v1_101
+        - resnet_v1_152
+        - resnet_v2_50
+        - resnet_v2_101
+        - resnet_v2_152
+        - nasnet_large
+        - nasnet_mobile
+        - pnasnet_large
+        - mobilenet_v2_100_224
+        - mobilenet_v2_130_224
+        - mobilenet_v2_140_224
+        - mobilenet_v3_small_100_224
+        - mobilenet_v3_small_075_224
+        - mobilenet_v3_large_100_224
+        - mobilenet_v3_large_075_224
+    dropout_rate (Optional[float]):
+        Fraction of input units to drop in the last layer. Value should be between 0.0 and 1.0.
+    trainable (Optional[bool]):
+        If true fine tuning will be performed on entire Hub model. If false only additional
+        layers will be trained.
+    l2_regularization_penalty (Optional[float]):
+        l2 regularization penalty.
+inputs:
+- {name: class_names, type: 'typing.List[str]', description: List of class names corresponding
+    to the input image data}
+- {name: model_name, type: String, description: Name of the TFHub model to load, default: efficientnetv2-xl-21k,
+  optional: true}
+- {name: dropout_rate, type: Float, description: Dropout rate, default: '0.2', optional: true}
+- name: trainable
+  type: Boolean
+  description: True if fine tuning should be performed
+  default: "True"
+  optional: true
+- {name: l2_regularization_penalty, type: Float, description: Regularization penalty,
+  default: '0.0001', optional: true}
+outputs:
+- {name: loaded_model_path, type: TensorflowSavedModel, description: Output path for
+    the loaded model}
+- {name: image_size_path, type: HeightWidth}
+implementation:
+  container:
+    image: us-docker.pkg.dev/vertex-ai/ready-to-go-image-classification/image-components:v0.1
+    # command is a list of strings (command-line arguments).
+    # The YAML language has two syntaxes for lists and you can use either of them.
+    # Here we use the "flow syntax" - comma-separated strings inside square brackets.
+    command: [
+      python3,
+      # Path of the program inside the container
+      /pipelines/component/src/loading_component.py,
+      --loaded-model-path,
+      {outputPath: loaded_model_path},
+      --class-names,
+      {inputValue: class_names},
+      --model-name,
+      {inputValue: model_name},
+      --dropout-rate,
+      {inputValue: dropout_rate},
+      --trainable,
+      {inputValue: trainable},
+      --l2-regularization-penalty,
+      {inputValue: l2_regularization_penalty},
+      --image-size-path,
+      {outputPath: image_size_path},
+    ]

community-content/pipeline_components/image_ml_model_training/pipeline.py

@@ -0,0 +1,62 @@
+# python3 -m pip install "kfp<2.0.0" "google-cloud-aiplatform>=1.16.0" --upgrade --quiet
+from kfp import components
+from kfp.v2 import dsl
+
+# %% Loading components
+upload_Tensorflow_model_to_Google_Cloud_Vertex_AI_op = components.load_component_from_url('https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/399405402d95f4a011e2d2e967c96f8508ba5688/community-content/pipeline_components/google-cloud/Vertex_AI/Models/Upload_Tensorflow_model/component.yaml')
+deploy_model_to_endpoint_op = components.load_component_from_url('https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/399405402d95f4a011e2d2e967c96f8508ba5688/community-content/pipeline_components/google-cloud/Vertex_AI/Models/Deploy_to_endpoint/component.yaml')
+transcode_imagedataset_tfrecord_from_csv_op = components.load_component_from_url('https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/main/community-content/pipeline_components/image_ml_model_training/transcode_tfrecord_image_dataset_from_csv/component.yaml')
+load_image_classification_model_from_tfhub_op = components.load_component_from_url('https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/b5b65198a6c2ffe8c0fa2aa70127e3325752df68/community-content/pipeline_components/image_ml_model_training/load_image_classification_model/component.yaml')
+preprocess_image_data_op = components.load_component_from_url('https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/main/community-content/pipeline_components/image_ml_model_training/preprocess_image_data/component.yaml')
+train_tensorflow_image_classification_model_op = components.load_component_from_url('https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/main/community-content/pipeline_components/image_ml_model_training/train_image_classification_model/component.yaml')
+
+
+# %% Pipeline definition
+def image_classification_pipeline():
+    class_names = ['daisy', 'dandelion', 'roses', 'sunflowers', 'tulips']
+    csv_image_data_path = 'gs://cloud-samples-data/ai-platform/flowers/flowers.csv'
+    deploy_model = False
+
+    image_data = dsl.importer(
+        artifact_uri=csv_image_data_path, artifact_class=dsl.Dataset).output
+
+    image_tfrecord_data = transcode_imagedataset_tfrecord_from_csv_op(
+        csv_image_data_path=image_data,
+        class_names=class_names
+    ).outputs['tfrecord_image_data_path']
+
+    loaded_model_outputs = load_image_classification_model_from_tfhub_op(
+        class_names=class_names,
+    ).outputs
+
+    preprocessed_data = preprocess_image_data_op(
+        image_tfrecord_data,
+        height_width_path=loaded_model_outputs['image_size_path'],
+    ).outputs
+
+    trained_model = (train_tensorflow_image_classification_model_op(
+        preprocessed_training_data_path = preprocessed_data['preprocessed_training_data_path'],
+        preprocessed_validation_data_path = preprocessed_data['preprocessed_validation_data_path'],
+        model_path=loaded_model_outputs['loaded_model_path']).
+                   set_cpu_limit('96').
+                   set_memory_limit('128G').
+                   add_node_selector_constraint('cloud.google.com/gke-accelerator', 'NVIDIA_TESLA_A100').
+                   set_gpu_limit('8').
+                   outputs['trained_model_path'])
+
+    vertex_model_name = upload_Tensorflow_model_to_Google_Cloud_Vertex_AI_op(
+        model=trained_model,
+    ).outputs['model_name']
+
+    # Deploying the model might incur additional costs over time
+    if deploy_model:
+        vertex_endpoint_name = deploy_model_to_endpoint_op(
+            model_name=vertex_model_name,
+        ).outputs['endpoint_name']
+
+pipeline_func = image_classification_pipeline
+
+# %% Pipeline submission
+if __name__ == '__main__':
+    from google.cloud import aiplatform
+    aiplatform.PipelineJob.from_pipeline_func(pipeline_func=pipeline_func).submit()

community-content/pipeline_components/image_ml_model_training/preprocess_image_data/component.yaml

@@ -0,0 +1,57 @@
+name: Preprocess image data
+description: |
+  Preprocess the image data and split between train and validation.
+  Args:
+    input_data_path (str):
+        Input path for the TFRecord image data. Data will be formatted as 'label' (encoded image
+        label), and 'image_raw' (the binary string of the image data).
+    height_width_path (str):
+        Path to square height and width to resize images to. File should contain single float value.
+        Value is dependent on training model.
+    preprocessed_training_data_path (str):
+        Output path for the TFRecord training data. Data will be formatted as 'label' (encoded image
+        label), and 'image_raw' (the binary string of the image data).
+    preprocessed_validation_data_path (str):
+        Output path for the TFRecord validation data. Data will be formatted as 'label' (encoded
+        image label), and 'image_raw' (the binary string of the image data).
+    validation_split (Optional[float]):
+        Fraction of data that will make up validation dataset. Value should be between 0.0 and 1.0.
+    seed (Optional[int]):
+        The global random seed to ensure the system gets a unique random sequence
+        that is deterministic (https://www.tensorflow.org/api_docs/python/tf/random/set_seed).
+inputs:
+- {name: input_data_path, type: ImageDatasetTFRecord, description: 'Input path for
+    the TFRecord image data,'}
+- {name: height_width_path, type: HeightWidth, description: 'Path to square height and width to
+    resize images to,'}
+- {name: validation_split, type: Float, description: 'Fraction of data that will make
+    up validation dataset,', default: '0.2', optional: true}
+- {name: seed, type: Integer, description: Random seed, default: '0', optional: true}
+outputs:
+- {name: preprocessed_training_data_path, type: ImageDatasetTFRecord, description: 'Output
+    path for the training data,'}
+- {name: preprocessed_validation_data_path, type: ImageDatasetTFRecord, description: 'Output
+    path for the validation data,'}
+implementation:
+  container:
+    image: us-docker.pkg.dev/vertex-ai/ready-to-go-image-classification/image-components:v0.1
+    # command is a list of strings (command-line arguments).
+    # The YAML language has two syntaxes for lists and you can use either of them.
+    # Here we use the "flow syntax" - comma-separated strings inside square brackets.
+    command: [
+      python3,
+      # Path of the program inside the container
+      /pipelines/component/src/preprocessing_component.py,
+      --input-data-path,
+      {inputPath: input_data_path},
+      --height-width-path,
+      {inputPath: height_width_path},
+      --validation-split,
+      {inputValue: validation_split},
+      --seed,
+      {inputValue: seed},
+      --preprocessed-training-data-path,
+      {outputPath: preprocessed_training_data_path},
+      --preprocessed-validation-data-path,
+      {outputPath: preprocessed_validation_data_path},
+    ]

community-content/pipeline_components/image_ml_model_training/train_image_classification_model/component.yaml

@@ -0,0 +1,90 @@
+name: Train tensorflow image classification model
+description: |
+  Creates a trained image classification TensorFlow model.
+  Args:
+    preprocessed_training_data_path (str):
+        Input path to the TFRecord training data. Data will be formatted as 'label' (encoded image
+        label), and 'image_raw' (the binary string of the image data).
+    preprocessed_validation_data_path (str):
+        Input path to the TFRecord validation data. Data will be formatted as 'label' (encoded
+        image label), and 'image_raw' (the binary string of the image data).
+    model_path (str):
+        Input path to the loaded pre-trained model.
+    trained_model_path (str):
+        Output path to save the trained model to.
+    optimizer_name (Optional[str]):
+        Name of the tf.keras optimizer. Available optimizers are listed at
+        https://keras.io/api/optimizers/
+    optimizer_parameters (Optional[Dict[str, str]]):
+        Optimizer parameters.
+    loss_function_name (Optional[str]):
+        Name of the loss function.
+    loss_function_parameters (Optional[Dict[str, str]]):
+        Loss function parameters.
+    number_of_epochs (Optional[int]):
+        Number of training iterations over data.
+    metric_names (Optional[Sequence[str]]):
+        List of tf.keras.metrics to be evaluated by the model during training and testing. Available
+        metrics are listed at https://keras.io/api/metrics/.
+    seed Optional(int):
+        The global random seed to ensure the system gets a unique random sequence
+        that is deterministic (https://www.tensorflow.org/api_docs/python/tf/random/set_seed).
+inputs:
+- {name: preprocessed_training_data_path, type: ImageDatasetTFRecord, description: 'Input
+    path for the training data,'}
+- {name: preprocessed_validation_data_path, type: ImageDatasetTFRecord, description: 'Input
+    path for the validation data,'}
+- {name: model_path, type: TensorflowSavedModel, description: 'Input path for the
+    model,'}
+- {name: optimizer_name, type: String, description: 'Name of the optimizer,', default: SGD,
+  optional: true}
+- {name: optimizer_parameters, type: 'typing.Dict[str, str]', description: 'Optimizer
+    parameters,', default: '{}', optional: true}
+- {name: loss_function_name, type: String, description: 'Name of the loss function,',
+  default: CategoricalCrossentropy, optional: true}
+- {name: loss_function_parameters, type: 'typing.Dict[str, str]', description: 'Loss
+    function parameters,', default: '{}', optional: true}
+- {name: number_of_epochs, type: Integer, description: 'Number of epochs,', default: '10',
+  optional: true}
+- {name: metric_names, type: 'typing.List[str]', description: 'List of metrics to
+    use,', default: '["accuracy"]', optional: true}
+- {name: seed, type: Integer, description: 'Random seed,', default: '0', optional: true}
+- {name: batch_size, type: Integer, description: Batch size, default: '16', optional: true}
+outputs:
+- {name: trained_model_path, type: TensorflowSavedModel, description: 'Output path
+    for the saved model,'}
+implementation:
+  container:
+    image: us-docker.pkg.dev/vertex-ai/ready-to-go-image-classification/image-components:v0.1
+    # command is a list of strings (command-line arguments).
+    # The YAML language has two syntaxes for lists and you can use either of them.
+    # Here we use the "flow syntax" - comma-separated strings inside square brackets.
+    command: [
+      python3,
+      # Path of the program inside the container
+      /pipelines/component/src/training_component.py,
+      --preprocessed-training-data-path,
+      {inputPath: preprocessed_training_data_path},
+      --preprocessed-validation-data-path,
+      {inputPath: preprocessed_validation_data_path},
+      --model-path,
+      {inputPath: model_path},
+      --trained-model-path,
+      {outputPath: trained_model_path},
+      --optimizer-name,
+      {inputValue: optimizer_name},
+      --loss-function-name,
+      {inputValue: loss_function_name},
+      --number-of-epochs,
+      {inputValue: number_of_epochs},
+      --seed,
+      {inputValue: seed},
+      --batch-size,
+      {inputValue: batch_size},
+      --metric-names,
+      {inputValue: metric_names},
+      --optimizer-parameters,
+      {inputValue: optimizer_parameters},
+      --loss-function-parameters,
+      {inputValue: loss_function_parameters},
+    ]

community-content/pipeline_components/image_ml_model_training/transcode_tfrecord_image_dataset_from_csv/component.yaml

@@ -0,0 +1,37 @@
+name: Transcode imagedataset tfrecord from csv
+description: |
+  Transcodes CSV Data into TFRecord file of TFExamples.
+  Args:
+    csv_image_data_path (str):
+        Path to the CSV image data. Data must include 'image_filepath' (Path to image file) and
+        'image_label' (output for a prediction) fields.
+    class_names (Sequence[str]):
+        Sequence of strings of categories for classification corresponding to input data.
+    tfrecord_image_data_path (str):
+        Output path for the TFRecord image data. Data will be formatted as 'label' (encoded image
+        label), and 'image_raw' (the binary string of the image data).
+inputs:
+- {name: csv_image_data_path, type: ImageDatasetCSV, description: Input path for the
+    CSV image data}
+- {name: class_names, type: 'typing.List[str]', description: List of class names corresponding
+    to the input image data}
+outputs:
+- {name: tfrecord_image_data_path, type: ImageDatasetTFRecord, description: Output
+    path for the TFRecord image data}
+implementation:
+  container:
+    image: us-docker.pkg.dev/vertex-ai/ready-to-go-image-classification/image-components:v0.1
+    # command is a list of strings (command-line arguments).
+    # The YAML language has two syntaxes for lists and you can use either of them.
+    # Here we use the "flow syntax" - comma-separated strings inside square brackets.
+    command: [
+      python3,
+      # Path of the program inside the container
+      /pipelines/component/src/transcoding_csv_component.py,
+      --csv-image-data-path,
+      {inputPath: csv_image_data_path},
+      --tfrecord-image-data-path,
+      {outputPath: tfrecord_image_data_path},
+      --class-names,
+      {inputValue: class_names},
+    ]

community-content/pipeline_components/image_ml_model_training/transcode_tfrecord_image_dataset_from_jsonl/component.yaml

@@ -0,0 +1,39 @@
+name: Transcode imagedataset tfrecord from jsonlines
+description: |
+  Transcodes JSONL Data into TFRecord file of TFExamples.
+  Args:
+    jsonl_image_data_path (str):
+        Input path for the JSONL image data
+        Path to the JSONL image data. Each line corresponds to a JSON input describing an image.
+        Schema follows AutoML image classification JSONL format
+        https://cloud.google.com/vertex-ai/docs/image-data/classification/prepare-data#json-lines.
+    class_names (Sequence[str]):
+        Sequence of strings of categories for classification corresponding to input data.
+    tfrecord_image_data_path (str):
+        Output path for the TFRecord image data. Data will be formatted as 'label' (encoded image
+        label), and 'image_raw' (the binary string of the image data).
+inputs:
+- {name: jsonl_image_data_path, type: ImageDatasetJsonLines, description: Input path
+    for the JSONL image data}
+- {name: class_names, type: 'typing.List[str]', description: List of class names corresponding
+    to the input image data}
+outputs:
+- {name: tfrecord_image_data_path, type: ImageDatasetTFRecord, description: Output
+    path for the TFRecord image data}
+implementation:
+  container:
+    image: us-docker.pkg.dev/vertex-ai/ready-to-go-image-classification/image-components:v0.1
+    # command is a list of strings (command-line arguments).
+    # The YAML language has two syntaxes for lists and you can use either of them.
+    # Here we use the "flow syntax" - comma-separated strings inside square brackets.
+    command: [
+      python3,
+      # Path of the program inside the container
+      /pipelines/component/src/transcoding_jsonl_component.py,
+      --jsonl-image-data-path,
+      {inputPath: jsonl_image_data_path},
+      --tfrecord-image-data-path,
+      {outputPath: tfrecord_image_data_path},
+      --class-names,
+      {inputValue: class_names},
+    ]

community-content/pytorch_efficient_training/README.md

@@ -15,15 +15,19 @@ pip install -r requirements.txt
 * resnet_dp.py - Train ResNet-50 on single node multiple GPUs with `DataParallel` strategy.
 * resnet_ddp.py - Train ResNet-50 on single node multiple GPUs with `DistributedDataParallel` strategy.
 * resnet_ddp_wds.py - Train ResNet-50 on single node multiple GPUs with `DistributedDataParallel` strategy and `Webdataset`.
+* resnet_fsdp.py - Train ResNet-50 on single node multiple GPUs with `FullyShardedDataParallel` strategy.
+* resnet_fsdp_wds.py - Train ResNet-50 on single node multiple GPUs with `FullyShardedDataParallel` strategy and `Webdataset`.
 * shard_imagenet.py - Shard ImagNet individual files into `tar` files.
 
 ## Benchmark
 
 When run the benchmark on Nvidia T4 GPUs using ImageNet validation dataset, you can get the result like:
-Strategy              | Seconds/Epoch - Local Data | Seconds/Epoch - Cloud Data
---------------------- | -------------------------- | --------------------------
-On 1 GPU              | 489                        | 804 (2x slower)
-On 4 GPUs (DP)        | 157                        | 738 (5x slower)
-On 4 GPUs (DDP)       | 134                        | 432 (3x slower)
-On 4 GPUs (DDP + WDS) | 131                        | 133 (same performance)
+Strategy               | Seconds/Epoch - Local Data | Seconds/Epoch - Cloud Data
+---------------------- | -------------------------- | --------------------------
+On 1 GPU               | 489                        | 804 (2x slower)
+On 4 GPUs (DP)         | 157                        | 738 (5x slower)
+On 4 GPUs (DDP)        | 134                        | 432 (3x slower)
+On 4 GPUs (DDP + WDS)  | 131                        | 133 (same performance)
+On 4 GPUs (FSDP)       | 139                        | 353 (3x slower)
+On 4 GPUs (FSDP + WDS) | 138                        | 135 (same performance)
 

community-content/pytorch_efficient_training/resnet_fsdp.py

@@ -0,0 +1,242 @@
+# Copyright 2022 Google LLC
+#
+# Licensed under the Apache License, Version 2.0 (the \"License\");
+# you may not use this file except in compliance with the License.\n",
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an \"AS IS\" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Train resnet on multiple GPUs with FSDP."""
+
+import argparse
+import functools
+import os
+import time
+
+from PIL import Image
+import torch
+from torch import nn
+import torch.distributed as dist
+from torch.distributed.fsdp import FullyShardedDataParallel as FSDP
+from torch.distributed.fsdp.wrap import size_based_auto_wrap_policy
+import torch.multiprocessing as mp
+import torchmetrics
+import torchvision
+from torchvision.models import resnet50
+
+
+class ImageFolder(torchvision.datasets.ImageFolder):
+  """Class for loading imagenet."""
+
+  def __init__(self, image_list_file, transform=None, target_transform=None):
+    self.samples = self._make_dataset(image_list_file)
+    self.loader = self._loader
+
+    self.imgs = self.samples
+    self.targets = [s[1] for s in self.samples]
+
+    self.transform = transform
+    self.target_transform = target_transform
+
+  def _make_dataset(self, image_list_file):
+    items = []
+    with open(image_list_file, 'r') as f:
+      for line in f:
+        item = line.strip().split(' ')
+        items.append((item[0], int(item[1])))
+    return items
+
+  def _loader(self, image_path):
+    with open(image_path, 'rb') as f:
+      img = Image.open(f)
+      img = img.convert('RGB')
+      return img
+
+
+def train(model, device, dataloader, optimizer):
+  model.train()
+  for image, target in dataloader:
+    image = image.to(device, non_blocking=True)
+    target = target.to(device, non_blocking=True)
+    pred = model(image)
+    # pred.shape (N, C), target.shape (N)
+    loss = nn.functional.cross_entropy(pred, target)
+    optimizer.zero_grad()
+    loss.backward()
+    optimizer.step()
+  return loss
+
+
+def evaluate(model, device, dataloader, metric):
+  model.eval()
+  with torch.no_grad():
+    for image, target in dataloader:
+      image = image.to(device, non_blocking=True)
+      target = target.to(device, non_blocking=True)
+      pred = model(image)
+      metric.update(pred, target)
+  accuracy = metric.compute()
+  metric.reset()
+  return accuracy
+
+
+def worker(gpu, args):
+  """Run training and evaluation."""
+  # Init process group.
+  print(f'Initiating process {gpu}')
+  dist.init_process_group(
+      backend='nccl',
+      init_method='env://',
+      world_size=args.gpus,
+      rank=gpu)
+
+  # Create train dataloader.
+  train_dataset = ImageFolder(
+      image_list_file=args.train_data_path,
+      transform=torchvision.transforms.Compose([
+          torchvision.transforms.RandomResizedCrop(224),
+          torchvision.transforms.RandomHorizontalFlip(),
+          torchvision.transforms.ToTensor(),
+          torchvision.transforms.Normalize(
+              mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),
+      ]))
+  train_sampler = torch.utils.data.distributed.DistributedSampler(
+      train_dataset, num_replicas=args.gpus, rank=gpu)
+  train_dataloader = torch.utils.data.DataLoader(
+      dataset=train_dataset,
+      batch_size=args.train_batch_size,
+      shuffle=False,
+      num_workers=args.dataloader_num_workers,
+      pin_memory=True,
+      sampler=train_sampler)
+  if gpu == 0:
+    print(f'Train dataloader | samples: {len(train_dataloader.dataset)}, '
+          f'num workers: {train_dataloader.num_workers}, '
+          f'global batch size: {args.train_batch_size * args.gpus}, '
+          f'batches/epoch: {len(train_dataloader)}')
+
+  # Create eval dataloader.
+  eval_dataset = ImageFolder(
+      image_list_file=args.eval_data_path,
+      transform=torchvision.transforms.Compose([
+          torchvision.transforms.Resize(256),
+          torchvision.transforms.CenterCrop(224),
+          torchvision.transforms.ToTensor(),
+          torchvision.transforms.Normalize(
+              mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),
+      ]))
+  eval_sampler = torch.utils.data.distributed.DistributedSampler(
+      eval_dataset, num_replicas=args.gpus, rank=gpu)
+  eval_dataloader = torch.utils.data.DataLoader(
+      dataset=eval_dataset,
+      batch_size=args.eval_batch_size,
+      shuffle=False,
+      num_workers=args.dataloader_num_workers,
+      pin_memory=True,
+      drop_last=True,
+      sampler=eval_sampler)
+  if gpu == 0:
+    print(f'Eval dataloader | samples: {len(eval_dataloader.dataset)}, '
+          f'num workers: {eval_dataloader.num_workers}, '
+          f'batch size: {args.eval_batch_size}, '
+          f'batches/epoch: {len(eval_dataloader)}')
+
+  # Wrap policy.
+  my_auto_wrap_policy = functools.partial(
+      size_based_auto_wrap_policy, min_num_params=100)
+  torch.cuda.set_device(gpu)
+
+  # Create model.
+  model = resnet50(weights=None)
+  model.to(args.device)
+  model = FSDP(model, auto_wrap_policy=my_auto_wrap_policy)
+
+  # Optimizer.
+  optimizer = torch.optim.SGD(model.parameters(), 0.1)
+
+  # Main loop.
+  metric = torchmetrics.classification.Accuracy(top_k=1).to(args.device)
+  for epoch in range(1, args.epochs + 1):
+    if gpu == 0:
+      print(f'Running epoch {epoch}')
+    train_sampler.set_epoch(epoch)
+
+    start = time.time()
+    train(model, args.device, train_dataloader, optimizer)
+    end = time.time()
+    if gpu == 0:
+      print(f'Training finished in {(end - start):>0.3f} seconds')
+
+    start = time.time()
+    evaluate(model, args.device, eval_dataloader, metric)
+    end = time.time()
+    if gpu == 0:
+      print(f'Evaluation finished in {(end - start):>0.3f} seconds')
+
+  if gpu == 0:
+    print('Done')
+  dist.destroy_process_group()
+
+
+def create_args():
+  """Create main args."""
+  parser = argparse.ArgumentParser(
+      formatter_class=argparse.ArgumentDefaultsHelpFormatter)
+  parser.add_argument(
+      '--gpus',
+      default=4,
+      type=int,
+      help='number of gpus to use')
+  parser.add_argument(
+      '--epochs',
+      default=2,
+      type=int,
+      help='number of total epochs to run')
+  parser.add_argument(
+      '--dataloader_num_workers',
+      default=2,
+      type=int,
+      help='number of workders for dataloader')
+  parser.add_argument(
+      '--train_data_path',
+      default='',
+      type=str,
+      help='path to training data')
+  parser.add_argument(
+      '--train_batch_size',
+      default=32,
+      type=int,
+      help='batch size for training per gpu')
+  parser.add_argument(
+      '--eval_data_path',
+      default='',
+      type=str,
+      help='path to evaluation data')
+  parser.add_argument(
+      '--eval_batch_size',
+      default=32,
+      type=int,
+      help='batch size for evaluation per gpu')
+  args = parser.parse_args()
+  return args
+
+
+def main():
+  args = create_args()
+
+  os.environ['MASTER_ADDR'] = 'localhost'
+  os.environ['MASTER_PORT'] = '8888'
+
+  args.device = 'cuda' if torch.cuda.is_available() else 'cpu'
+  print(f'Launch job on {args.gpus} GPUs with FSDP')
+  mp.spawn(worker, nprocs=args.gpus, args=(args,))
+
+
+if __name__ == '__main__':
+  main()

community-content/pytorch_efficient_training/resnet_fsdp_wds.py

@@ -0,0 +1,240 @@
+"""Train resnet on multiple GPUs with DDP."""
+
+import argparse
+import functools
+import itertools
+import math
+import os
+import time
+
+import torch
+from torch import nn
+import torch.distributed as dist
+from torch.distributed.fsdp import FullyShardedDataParallel as FSDP
+from torch.distributed.fsdp.wrap import size_based_auto_wrap_policy
+import torch.multiprocessing as mp
+import torchmetrics
+from torchvision.models import resnet50
+from torchvision.transforms import transforms
+import webdataset as wds
+
+
+def wds_split(src, rank, world_size):
+  """Shards split function for webdataset."""
+  # The context of caller of this function is within multiple processes
+  # (by DDP world_size) and multiple workers (by dataloader_num_workers).
+  # So we totally have (world_size * num_workers) workers for processing data.
+  # NOTE: Raw data should be sharded to enough shards to make sure one process
+  # can handle at least one shard, otherwise the process may hang.
+  worker_id = 0
+  num_workers = 1
+  worker_info = torch.utils.data.get_worker_info()
+  if worker_info:
+    worker_id = worker_info.id
+    num_workers = worker_info.num_workers
+  for s in itertools.islice(src, rank * num_workers + worker_id, None,
+                            world_size * num_workers):
+    yield s
+
+
+def identity(x):
+  return x
+
+
+def create_wds_dataloader(rank, args, mode):
+  """Create webdataset dataset and dataloader."""
+  if mode == 'train':
+    transform = transforms.Compose([
+        transforms.RandomResizedCrop(224),
+        transforms.RandomHorizontalFlip(),
+        transforms.ToTensor(),
+        transforms.Normalize(
+            mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),
+    ])
+    data_path = args.train_data_path
+    data_size = args.train_data_size
+    batch_size_local = args.train_batch_size
+    batch_size_global = args.train_batch_size * args.gpus
+    # Since webdataset disallows partial batch, we pad the last batch for train.
+    batches = int(math.ceil(data_size / batch_size_global))
+  else:
+    transform = transforms.Compose([
+        transforms.Resize(256),
+        transforms.CenterCrop(224),
+        transforms.ToTensor(),
+        transforms.Normalize(
+            mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),
+    ])
+    data_path = args.eval_data_path
+    data_size = args.eval_data_size
+    batch_size_local = args.eval_batch_size
+    batch_size_global = args.eval_batch_size * args.gpus
+    # Since webdataset disallows partial batch, we drop the last batch for eval.
+    batches = int(data_size / batch_size_global)
+
+  dataset = wds.DataPipeline(
+      wds.SimpleShardList(data_path),
+      functools.partial(wds_split, rank=rank, world_size=args.gpus),
+      wds.tarfile_to_samples(),
+      wds.decode('pil'),
+      wds.to_tuple('jpg;png;jpeg cls'),
+      wds.map_tuple(transform, identity),
+      wds.batched(batch_size_local, partial=False),
+  )
+  num_workers = args.dataloader_num_workers
+  dataloader = wds.WebLoader(
+      dataset=dataset,
+      batch_size=None,
+      shuffle=False,
+      num_workers=num_workers,
+      persistent_workers=True if num_workers > 0 else False,
+      pin_memory=True).repeat(nbatches=batches)
+  print(f'{mode} dataloader | samples: {data_size}, '
+        f'num_workers: {num_workers}, '
+        f'local batch size: {batch_size_local}, '
+        f'global batch size: {batch_size_global}, '
+        f'batches: {batches}')
+  return dataloader
+
+
+def train(model, device, dataloader, optimizer):
+  model.train()
+  for image, target in dataloader:
+    image = image.to(device, non_blocking=True)
+    target = target.to(device, non_blocking=True)
+    pred = model(image)
+    # pred.shape (N, C), target.shape (N)
+    loss = nn.functional.cross_entropy(pred, target)
+    optimizer.zero_grad()
+    loss.backward()
+    optimizer.step()
+  return loss
+
+
+def evaluate(model, device, dataloader, metric):
+  model.eval()
+  with torch.no_grad():
+    for image, target in dataloader:
+      image = image.to(device, non_blocking=True)
+      target = target.to(device, non_blocking=True)
+      pred = model(image)
+      metric.update(pred, target)
+  accuracy = metric.compute()
+  metric.reset()
+  return accuracy
+
+
+def worker(gpu, args):
+  """Run training and evaluation."""
+  # Init process group.
+  print(f'Initiating process {gpu}')
+  dist.init_process_group(
+      backend='nccl',
+      init_method='env://',
+      world_size=args.gpus,
+      rank=gpu)
+
+  # Create dataloader.
+  train_dataloader = create_wds_dataloader(gpu, args, 'train')
+  eval_dataloader = create_wds_dataloader(gpu, args, 'eval')
+
+  # Wrap policy.
+  my_auto_wrap_policy = functools.partial(
+      size_based_auto_wrap_policy, min_num_params=100)
+  torch.cuda.set_device(gpu)
+
+  # Create model.
+  model = resnet50(weights=None)
+  model.to(args.device)
+  model = FSDP(model, auto_wrap_policy=my_auto_wrap_policy)
+
+  # Optimizer.
+  optimizer = torch.optim.SGD(model.parameters(), 0.1)
+
+  # Main loop.
+  metric = torchmetrics.classification.Accuracy(top_k=1).to(args.device)
+  for epoch in range(1, args.epochs + 1):
+    if gpu == 0:
+      print(f'Running epoch {epoch}')
+
+    start = time.time()
+    train(model, args.device, train_dataloader, optimizer)
+    end = time.time()
+    if gpu == 0:
+      print(f'Training finished in {(end - start):>0.3f} seconds')
+
+    start = time.time()
+    evaluate(model, args.device, eval_dataloader, metric)
+    end = time.time()
+    if gpu == 0:
+      print(f'Evaluation finished in {(end - start):>0.3f} seconds')
+
+  if gpu == 0:
+    print('Done')
+
+
+def create_args():
+  """Create main args."""
+  parser = argparse.ArgumentParser(
+      formatter_class=argparse.ArgumentDefaultsHelpFormatter)
+  parser.add_argument(
+      '--gpus',
+      default=4,
+      type=int,
+      help='number of gpus to use')
+  parser.add_argument(
+      '--epochs',
+      default=2,
+      type=int,
+      help='number of total epochs to run')
+  parser.add_argument(
+      '--dataloader_num_workers',
+      default=2,
+      type=int,
+      help='number of workders for dataloader')
+  parser.add_argument(
+      '--train_data_path',
+      default='',
+      type=str,
+      help='path to training data')
+  parser.add_argument(
+      '--train_batch_size',
+      default=32,
+      type=int,
+      help='batch size for training per gpu')
+  parser.add_argument(
+      '--train_data_size',
+      default=50000,
+      type=int,
+      help='data size for training')
+  parser.add_argument(
+      '--eval_data_path',
+      default='',
+      type=str,
+      help='path to evaluation data')
+  parser.add_argument(
+      '--eval_batch_size',
+      default=32,
+      type=int,
+      help='batch size for evaluation per gpu')
+  parser.add_argument(
+      '--eval_data_size',
+      default=50000,
+      type=int,
+      help='data size for evaluation')
+  args = parser.parse_args()
+  return args
+
+
+def main():
+  args = create_args()
+  os.environ['MASTER_ADDR'] = 'localhost'
+  os.environ['MASTER_PORT'] = '8888'
+
+  args.device = 'cuda' if torch.cuda.is_available() else 'cpu'
+  print(f'Launch job on {args.gpus} GPUs with FSDP')
+  mp.spawn(worker, nprocs=args.gpus, args=(args,))
+
+
+if __name__ == '__main__':
+  main()

community-content/pytorch_image_classification_distributed_data_parallel_training_with_vertex_sdk/trainer/requirements.txt

@@ -1,3 +1,3 @@
-torch==1.8.1
+torch==1.13.1
 torchvision==0.9.1
 tensorboard==2.5.0

community-content/pytorch_image_classification_single_gpu_with_vertex_sdk_and_torchserve/trainer/requirements.txt

@@ -1,3 +1,3 @@
-torch==1.8.1
+torch==1.13.1
 torchvision==0.9.1
 tensorboard==2.5.0

community-content/pytorch_pre_built_images_deployment/pytorch-text-classification-vertex-ai-deploy.ipynb

@@ -31,17 +31,7 @@
       "source": [
         "# Deploying a PyTorch Text Classification Model on [Vertex AI](https://cloud.google.com/vertex-ai)\n",
         "\n",
-        "**This is an Experimental release**, covered by the Pre-GA Offerings Terms of your Google Cloud Platform [Terms of Service](https://cloud.google.com/terms).\n",
-        "\n",
-        "Experiments are focused on validating a prototype and are not guaranteed to be released. They are not intended for production use or covered by any SLA, support obligation, or deprecation policy and might be subject to backward-incompatible changes.\n",
-        "\n",
-        "**Kindly drop us a note before you run any scale tests.**\n",
-        "\n",
-        "**Do not hesitate to contact vertexai-prediction-preview-feedback@google.com if you have any questions or run into any issues.**\n",
-        "\n",
-        "The usage of the product is free during the Experimental release period: you will still incur charges for other GCP products usage, such as storage.\n",
-        "\n",
-        "The projects need to be added to the allowlist in order to deploy PyTorch models using Vertex AI Prediction pre-built PyTorch images. If you are interested in the feature, please send an email to vertexai-prediction-preview-feedback@google.com to provide your project numbers OR project ids."
+        "**Kindly reach out to Vertex AI before you run any scale tests or you have any questions.**\n"
       ]
     },
     {

notebooks/community/CODEOWNERS

@@ -30,6 +30,7 @@
 /notebooks/community/neo4j/graph_paysim.ipynb @benofben @laeg
 /notebooks/community/ml_ops/stage1/get_started_with_visionapi_and_vertex_datasets.ipynb @mansari
 /notebooks/community/pipelines/google_cloud_pipeline_components_bqml_pipeline_demand_forecasting.ipynb @inardini
+/notebooks/community/cohere/cohere_embedding_with_matching_engine.ipynb @stewart-co
 /notebooks/community/ml_ops/stage2/get_started_vertex_hpt_r_kernel.ipynb @fhirschmann
 /notebooks/community/ml_ops/stage2/get_started_vertex_training_r_using_r_kernel.ipynb @fhirschmann
 /notebooks/community/vertex-ai-samples/notebooks/community/model_registry/vertex_ai_model_registry_bqml_custom_model_versioning.ipynb @inardini
@@ -37,3 +38,20 @@
 /notebooks/community/vizier/conversions_vertex_vizier_and_open_source_vizier.ipynb @halio-g
 /notebooks/community/experiments/vertex_ai_model_experimentation.ipynb @inardini @asobran
 /notebooks/community/pipelines/google_cloud_pipeline_components_bqml_pipeline_anomaly_detection.ipynb @inardini
+/notebooks/community/pipelines/google_cloud_pipeline_components_cloud_natural_language_pipeline.ipynb @Narwhalprime
+/notebooks/community/pipelines/google_cloud_pipeline_components_ready_to_go_text_classification_pipeline.ipynb @Narwhalprime
+/notebooks/community/feature_store/get_started_vertex_feature_store.ipynb @junkourata
+/notebooks/community/model_garden/model_garden_tfvision_image_classification.ipynb @genquan9
+/notebooks/community/model_garden/model_garden_tfvision_image_object_detection.ipynb @genquan9
+/notebooks/community/model_garden/model_garden_tfvision_image_segmentation.ipynb @genquan9
+/notebooks/community/model_garden/model_garden_pytorch_stable_diffusion.ipynb @xiangxu-google
+/notebooks/community/model_garden/model_garden_pytorch_stable_diffusion_inpainting.ipynb @xiangxu-google
+/notebooks/community/model_garden/model_garden_pytorch_instructpix2pix.ipynb @xiangxu-google
+/notebooks/community/model_garden/model_garden_pytorch_controlnet.ipynb @xiangxu-google
+/notebooks/community/model_garden/model_garden_pytorch_blip_image_captioning.ipynb @xiangxu-google
+/notebooks/community/model_garden/model_garden_pytorch_blip_vqa.ipynb @xiangxu-google
+/notebooks/community/model_garden/model_garden_pytorch_vilt_vqa.ipynb @xiangxu-google
+/notebooks/community/model_garden/model_garden_pytorch_vit_gpt2_image_captioning.ipynb @xiangxu-google
+/notebooks/community/model_garden/model_garden_pytorch_clip.ipynb @xiangxu-google
+/notebooks/community/model_garden/model_garden_pytorch_owlvit.ipynb @xiangxu-google
+/notebooks/community/model_garden/model_garden_pytorch_layoutml_document_qa.ipynb @xiangxu-google

notebooks/community/cohere/README.md

@@ -0,0 +1,3 @@
+# README
+
+These are notebooks [Cohere](https://cohere.ai/) built in collaboration with Google.  They demonstrate how to use Cohere's modeling API along with Vertex AI.

notebooks/community/matching_engine/matching_engine_for_indexing.ipynb

@@ -8,7 +8,7 @@
       },
       "outputs": [],
       "source": [
-        "# Copyright 2021 Google LLC\n",
+        "# Copyright 2023 Google LLC\n",
         "#\n",
         "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
         "# you may not use this file except in compliance with the License.\n",
@@ -24,6 +24,7 @@
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "id": "JAPoU8Sm5E6e"
@@ -32,20 +33,28 @@
         "<table align=\"left\">\n",
         "\n",
         "  <td>\n",
-        "    <a href=\"https://console.cloud.google.com/vertex-ai/notebooks/deploy-notebook?download_url=https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/master/notebooks/community/matching_engine/matching_engine_for_indexing.ipynb\">\n",
-        "      Run in Google Cloud Notebooks\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/matching_engine/matching_engine_for_indexing.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\">\n",
+        "      Run in Colab\n",
         "    </a>\n",
         "  </td>\n",
         "  <td>\n",
-        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/master/notebooks/community/matching_engine/matching_engine_for_indexing.ipynb\">\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/matching_engine/matching_engine_for_indexing.ipynb\">\n",
         "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
         "      View on GitHub\n",
         "    </a>\n",
         "  </td>\n",
+        "  <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/workbench/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/main/notebooks/community/matching_engine/matching_engine_for_indexing.ipynb\">\n",
+        "      <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "      Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>                                                                                               \n",
         "</table>"
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "id": "tvgnzT1CKxrO"
@@ -53,25 +62,49 @@
       "source": [
         "## Overview\n",
         "\n",
-        "This example demonstrates how to use the GCP ANN Service. It is a high scale, low latency solution, to find similar vectors (or more specifically \"embeddings\") for a large corpus. Moreover, it is a fully managed offering, further reducing operational overhead. It is built upon [Approximate Nearest Neighbor (ANN) technology](https://ai.googleblog.com/2020/07/announcing-scann-efficient-vector.html) developed by Google Research.\n",
-        "\n",
-        "### Dataset\n",
-        "\n",
-        "The dataset used for this tutorial is the [GloVe dataset](https://nlp.stanford.edu/projects/glove/).\n",
-        "\n",
+        "This example demonstrates how to use Vertex AI Matching Engine. It is a high scale, low latency solution, to find similar vectors (or more specifically \"embeddings\") for a large corpus. Moreover, it is a fully managed offering, further reducing operational overhead. It is built upon [Approximate Nearest Neighbor (ANN) technology](https://ai.googleblog.com/2020/07/announcing-scann-efficient-vector.html) developed by Google Research."
+      ]
+    },
+    {
+      "attachments": {},
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "56e5f9699c6c"
+      },
+      "source": [
         "### Objective\n",
         "\n",
         "In this notebook, you will learn how to create Approximate Nearest Neighbor (ANN) Index, query against indexes, and validate the performance of the index. \n",
         "\n",
         "The steps performed include:\n",
         "\n",
-        "* Create ANN Index and Brute Force Index\n",
+        "* Create a Vertex AI Matching Engine Index and Brute Force Index\n",
         "* Create an IndexEndpoint with VPC Network\n",
-        "* Deploy ANN Index and Brute Force Index\n",
-        "* Perform online query\n",
-        "* Compute recall\n",
-        "\n",
+        "* Deploy a Vertex AI Matching Engine Index and Brute Force Index\n",
+        "* Perform online queries\n",
+        "* Submit batch queries\n",
+        "* Compute recall metric"
+      ]
+    },
+    {
+      "attachments": {},
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0aaef374550b"
+      },
+      "source": [
+        "### Dataset\n",
         "\n",
+        "The dataset used for this tutorial is the [GloVe dataset](https://nlp.stanford.edu/projects/glove/)."
+      ]
+    },
+    {
+      "attachments": {},
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "5e2eba58ad71"
+      },
+      "source": [
         "### Costs \n",
         "\n",
         "This tutorial uses billable components of Google Cloud:\n",
@@ -87,6 +120,7 @@
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "id": "S5zc4kbEiYCm"
@@ -94,79 +128,47 @@
       "source": [
         "## Before you begin\n",
         "\n",
-        "* **Prepare a VPC network**.  To reduce any network overhead that might lead to unnecessary increase in overhead latency, it is best to call the ANN endpoints from your VPC via a direct [VPC Peering](https://cloud.google.com/vertex-ai/docs/general/vpc-peering) connection. The following section describes how to setup a VPC Peering connection if you don't have one. This is a one-time initial setup task. You can also reuse existing VPC network and skip this section.\n",
-        "* **WARNING:** The match service gRPC API (to create online queries against your deployed index) has to be executed in a Google Cloud Notebook instance that is created with the following requirements:\n",
-        "  * **In the same region as where your ANN service is deployed** (for example, if you set `REGION = \"us-central1\"` as same as the tutorial, the notebook instance has to be in `us-central1`).\n",
-        "  * **Make sure you select the VPC network you created for ANN service** (instead of using the \"default\" one). That is, you will have to create the VPC network below and then create a new notebook instance that uses that VPC.  \n",
-        "  * If you run it in the colab or a Google Cloud Notebook instance in a different VPC network or region, the gRPC API will fail to peer the network (InactiveRPCError)."
-      ]
-    },
-    {
-      "cell_type": "code",
-      "execution_count": null,
-      "metadata": {
-        "id": "lW2LneA5mmmP"
-      },
-      "outputs": [],
-      "source": [
-        "PROJECT_ID = \"<your_project_id>\"  # @param {type:\"string\"}\n",
-        "NETWORK_NAME = \"ucaip-haystack-vpc-network\"  # @param {type:\"string\"}\n",
-        "PEERING_RANGE_NAME = \"ucaip-haystack-range\"\n",
+        "### Set up your Google Cloud project\n",
         "\n",
-        "# Create a VPC network\n",
-        "! gcloud compute networks create {NETWORK_NAME} --bgp-routing-mode=regional --subnet-mode=auto --project={PROJECT_ID}\n",
+        "**The following steps are required, regardless of your notebook environment.**\n",
         "\n",
-        "# Add necessary firewall rules\n",
-        "! gcloud compute firewall-rules create {NETWORK_NAME}-allow-icmp --network {NETWORK_NAME} --priority 65534 --project {PROJECT_ID} --allow icmp\n",
+        "1. [Select or create a Google Cloud project](https://console.cloud.google.com/cloud-resource-manager). When you first create an account, you get a $300 free credit towards your compute/storage costs.\n",
         "\n",
-        "! gcloud compute firewall-rules create {NETWORK_NAME}-allow-internal --network {NETWORK_NAME} --priority 65534 --project {PROJECT_ID} --allow all --source-ranges 10.128.0.0/9\n",
+        "2. [Make sure that billing is enabled for your project](https://cloud.google.com/billing/docs/how-to/modify-project).\n",
         "\n",
-        "! gcloud compute firewall-rules create {NETWORK_NAME}-allow-rdp --network {NETWORK_NAME} --priority 65534 --project {PROJECT_ID} --allow tcp:3389\n",
+        "3. [Enable the Vertex AI API](https://console.cloud.google.com/flows/enableapi?apiid=aiplatform.googleapis.com).\n",
         "\n",
-        "! gcloud compute firewall-rules create {NETWORK_NAME}-allow-ssh --network {NETWORK_NAME} --priority 65534 --project {PROJECT_ID} --allow tcp:22\n",
-        "\n",
-        "# Reserve IP range\n",
-        "! gcloud compute addresses create {PEERING_RANGE_NAME} --global --prefix-length=16 --network={NETWORK_NAME} --purpose=VPC_PEERING --project={PROJECT_ID} --description=\"peering range for uCAIP Haystack.\"\n",
-        "\n",
-        "# Set up peering with service networking\n",
-        "! gcloud services vpc-peerings connect --service=servicenetworking.googleapis.com --network={NETWORK_NAME} --ranges={PEERING_RANGE_NAME} --project={PROJECT_ID}"
+        "4. If you are running this notebook locally, you need to install the [Cloud SDK](https://cloud.google.com/sdk)."
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
-        "id": "d3uj8x73nDX_"
-      },
-      "source": [
-        "* Authentication: `$ gcloud auth login` rerun this in Google Cloud Notebook terminal when you are logged out and need the credential again."
-      ]
-    },
-    {
-      "cell_type": "markdown",
-      "metadata": {
-        "id": "i7EUnXsZhAGF"
+        "id": "4700b0e39c5d"
       },
       "source": [
         "### Installation\n",
         "\n",
-        "Download and install the latest (preview) version of the Vertex SDK for Python."
+        "Download and install the latest version of the Vertex AI SDK for Python."
       ]
     },
     {
       "cell_type": "code",
       "execution_count": null,
       "metadata": {
-        "id": "wyy5Lbnzg5fi"
+        "id": "014470c6a8de"
       },
       "outputs": [],
       "source": [
-        "! pip install -U git+https://github.com/googleapis/python-aiplatform.git@main-test --user"
+        "! pip install -U git+https://github.com/googleapis/python-aiplatform.git@main --user"
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
-        "id": "irSMQn6gZ19l"
+        "id": "cf00462144f7"
       },
       "source": [
         "Install the `h5py` to prepare sample dataset, and the `grpcio-tools` for querying against the index. "
@@ -176,11 +178,15 @@
       "cell_type": "code",
       "execution_count": null,
       "metadata": {
-        "id": "-h5sqwOEZ5Yq"
+        "id": "3f3e45e5a1d1"
       },
       "outputs": [],
       "source": [
-        "! pip install -U grpcio-tools --user\n",
+        "! pip install protobuf==3.20.*\n",
+        "! pip install -U google-api-python-client==1.8.0 --user\n",
+        "! pip install -U grpcio-tools==1.47.0 --user\n",
+        "! pip install -U grpcio==1.47.0 --user\n",
+        "! pip install -U grpcio-status==1.47.0 --user\n",
         "! pip install -U h5py --user"
       ]
     },
@@ -199,7 +205,7 @@
       "cell_type": "code",
       "execution_count": null,
       "metadata": {
-        "id": "EzrelQZ22IZj"
+        "id": "aa1d87bdc90b"
       },
       "outputs": [],
       "source": [
@@ -215,79 +221,216 @@
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
-        "id": "BF1j6f9HApxa"
+        "id": "249da91c1011"
       },
       "source": [
-        "### Set up your Google Cloud project\n",
+        "### Set your project ID\n",
         "\n",
-        "**The following steps are required, regardless of your notebook environment.**\n",
-        "\n",
-        "1. [Select or create a Google Cloud project](https://console.cloud.google.com/cloud-resource-manager).\n",
-        "\n",
-        "1. [Make sure that billing is enabled for your project](https://cloud.google.com/billing/docs/how-to/modify-project).\n",
-        "\n",
-        "1. [Enable the Vertex AI API and Compute Engine API, and Service Networking API](https://console.cloud.google.com/flows/enableapi?apiid=aiplatform.googleapis.com,compute_component,servicenetworking.googleapis.com).\n",
-        "\n",
-        "1. Enter your project ID in the cell below. Then run the cell to make sure the\n",
-        "Cloud SDK uses the right project for all the commands in this notebook.\n",
-        "\n",
-        "**Note**: Jupyter runs lines prefixed with `!` as shell commands, and it interpolates Python variables prefixed with `$` into these commands."
-      ]
-    },
-    {
-      "cell_type": "markdown",
-      "metadata": {
-        "id": "WReHDGG5g0XY"
-      },
-      "source": [
-        "#### Set your project ID\n",
-        "\n",
-        "**If you don't know your project ID**, you may be able to get your project ID using `gcloud`."
+        "**If you don't know your project ID**, try the following:\n",
+        "* Run `gcloud config list`.\n",
+        "* Run `gcloud projects list`.\n",
+        "* See the support page: [Locate the project ID](https://support.google.com/googleapi/answer/7014113)"
       ]
     },
     {
       "cell_type": "code",
       "execution_count": null,
       "metadata": {
-        "id": "oM1iC_MfAts1"
+        "id": "10e0d2ee8c45"
       },
       "outputs": [],
       "source": [
-        "import os\n",
+        "PROJECT_ID = \"[your-project-id]\"  # @param {type:\"string\"}\n",
         "\n",
-        "PROJECT_ID = \"\"\n",
-        "\n",
-        "# Get your Google Cloud project ID from gcloud\n",
-        "if not os.getenv(\"IS_TESTING\"):\n",
-        "    shell_output=!gcloud config list --format 'value(core.project)' 2>/dev/null\n",
-        "    PROJECT_ID = shell_output[0]\n",
-        "    print(\"Project ID: \", PROJECT_ID)"
+        "# Set the project id\n",
+        "! gcloud config set project {PROJECT_ID}"
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
-        "id": "qJYoRfYng0XZ"
+        "id": "3fbfae3ff12a"
       },
       "source": [
-        "Otherwise, set your project ID here."
+        "### Set the region\n",
+        "\n",
+        "You can also change the `REGION` variable used by Vertex AI. Learn more about [Vertex AI regions](https://cloud.google.com/vertex-ai/docs/general/locations).\n",
+        "* **WARNING:** \n",
+        "    *  **Make sure to [choose a region where Vertex AI services are available](https://cloud.google.com/vertex-ai/docs/general/locations#available_regions).**\n",
+        "    *  **If you use Vertex Workbench, the Notebook instance needs to be in the same region where your Vertex AI Matching Engine is deployed.** (for example, if you set `REGION = \"us-central1\"` as same as the tutorial, the notebook instance has to be in `us-central1`)."
       ]
     },
     {
       "cell_type": "code",
       "execution_count": null,
       "metadata": {
-        "id": "riG_qUokg0XZ"
+        "id": "71c3fd82024e"
       },
       "outputs": [],
       "source": [
-        "if PROJECT_ID == \"\" or PROJECT_ID is None:\n",
-        "    PROJECT_ID = \"<your_project_id>\"  # @param {type:\"string\"}"
+        "REGION = \"us-central1\"  # @param {type: \"string\"}\n",
+        "\n",
+        "# Set the regions\n",
+        "! gcloud config set ai_platform/region {REGION}"
       ]
     },
     {
+      "attachments": {},
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "60c5a0f69ad8"
+      },
+      "source": [
+        "### Authenticate your Google Cloud account\n",
+        "\n",
+        "Depending on your Jupyter environment, you may have to manually authenticate. Follow the relevant instructions below."
+      ]
+    },
+    {
+      "attachments": {},
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d118c95af93f"
+      },
+      "source": [
+        "**1. Vertex AI Workbench**\n",
+        "* Do nothing as you are already authenticated."
+      ]
+    },
+    {
+      "attachments": {},
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "3035286fcdda"
+      },
+      "source": [
+        "**2. Local JupyterLab instance, uncomment and run:**"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "455882ec0f11"
+      },
+      "outputs": [],
+      "source": [
+        "# ! gcloud auth login"
+      ]
+    },
+    {
+      "attachments": {},
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "5097f3233d53"
+      },
+      "source": [
+        "**3. Colab, uncomment and run:**"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "2b88e46ac2c8"
+      },
+      "outputs": [],
+      "source": [
+        "# from google.colab import auth\n",
+        "# auth.authenticate_user()"
+      ]
+    },
+    {
+      "attachments": {},
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "fcdbb8929927"
+      },
+      "source": [
+        "**4. Service account or other**\n",
+        "* See how to grant Cloud Storage permissions to your service account at https://cloud.google.com/storage/docs/gsutil/commands/iam#ch-examples."
+      ]
+    },
+    {
+      "attachments": {},
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "7c6eef70dfdb"
+      },
+      "source": [
+        "### Prepare a VPC network\n",
+        "\n",
+        "To reduce any network overhead that might lead to unnecessary increase in overhead latency, it is best to call the Vertex AI Matching Engine endpoints from your VPC via a direct [VPC Peering](https://cloud.google.com/vertex-ai/docs/general/vpc-peering) connection. The following section describes how to setup a VPC Peering connection if you don't have one. This is a one-time initial setup task. You can also reuse existing VPC network and skip this section.\n",
+        "\n",
+        "* **WARNING:** The match service gRPC API (to create online queries against your deployed index) has to be executed in a Google Cloud Notebook instance that is created with the following requirements:\n",
+        "    * **Make sure you select the VPC network you created for Vertex AI Matching Engine service** (instead of using the \"default\" one). That is, you will have to create the VPC network below and then create a new notebook instance that uses that VPC.  \n",
+        "    * If you run it in the colab or a Google Cloud Notebook instance in a different VPC network or region, the gRPC API will fail to peer the network (InactiveRPCError)."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "ab38a8cc634c"
+      },
+      "outputs": [],
+      "source": [
+        "NETWORK_NAME = \"ucaip-haystack-vpc-network\"  # @param {type:\"string\"}\n",
+        "PEERING_RANGE_NAME = \"ucaip-haystack-range\""
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "ec6bf3199835"
+      },
+      "outputs": [],
+      "source": [
+        "# Create a VPC network\n",
+        "! gcloud compute networks create {NETWORK_NAME} --bgp-routing-mode=regional --subnet-mode=auto --project={PROJECT_ID}\n",
+        "\n",
+        "# Add necessary firewall rules\n",
+        "! gcloud compute firewall-rules create {NETWORK_NAME}-allow-icmp --network {NETWORK_NAME} --priority 65534 --project {PROJECT_ID} --allow icmp\n",
+        "\n",
+        "! gcloud compute firewall-rules create {NETWORK_NAME}-allow-internal --network {NETWORK_NAME} --priority 65534 --project {PROJECT_ID} --allow all --source-ranges 10.128.0.0/9\n",
+        "\n",
+        "! gcloud compute firewall-rules create {NETWORK_NAME}-allow-rdp --network {NETWORK_NAME} --priority 65534 --project {PROJECT_ID} --allow tcp:3389\n",
+        "\n",
+        "! gcloud compute firewall-rules create {NETWORK_NAME}-allow-ssh --network {NETWORK_NAME} --priority 65534 --project {PROJECT_ID} --allow tcp:22\n",
+        "\n",
+        "# Reserve IP range\n",
+        "! gcloud compute addresses create {PEERING_RANGE_NAME} --global --prefix-length=16 --network={NETWORK_NAME} --purpose=VPC_PEERING --project={PROJECT_ID} --description=\"peering range for uCAIP Haystack.\""
+      ]
+    },
+    {
+      "attachments": {},
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "ddbace09fe81"
+      },
+      "source": [
+        "Create the VPC Peering. If you are running this from Vertex AI Workbench it is possible you might need your notebook's instance service or user account to have the Service Networking Admin Role"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "d329aa3c54d3"
+      },
+      "outputs": [],
+      "source": [
+        "# Set up peering with service networking\n",
+        "! gcloud services vpc-peerings connect --service=servicenetworking.googleapis.com --network={NETWORK_NAME} --ranges={PEERING_RANGE_NAME} --project={PROJECT_ID}"
+      ]
+    },
+    {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "id": "zgPO1eR3CYjk"
@@ -297,13 +440,11 @@
         "\n",
         "**The following steps are required, regardless of your notebook environment.**\n",
         "\n",
-        "Set the name of your Cloud Storage bucket below. It must be unique across all\n",
+        "Create a storage bucket to store intermediate artifacts such as datasets. Set the name of your Cloud Storage bucket below. It must be unique across all\n",
         "Cloud Storage buckets.\n",
         "\n",
-        "You may also change the `REGION` variable, which is used for operations\n",
-        "throughout the rest of this notebook. Make sure to [choose a region where Vertex AI services are\n",
-        "available](https://cloud.google.com/vertex-ai/docs/general/locations#available_regions). You may\n",
-        "not use a Multi-Regional Storage bucket for training with Vertex AI."
+        "* **WARNING:** \n",
+        "    * **You may not use a Multi-Regional Storage bucket for training with Vertex AI.**"
       ]
     },
     {
@@ -314,8 +455,7 @@
       },
       "outputs": [],
       "source": [
-        "BUCKET_NAME = \"gs://[your-bucket-name]\"  # @param {type:\"string\"}\n",
-        "REGION = \"us-central1\"  # @param {type:\"string\"}"
+        "BUCKET_NAME = \"gs://[your-bucket-name-unique]\"  # @param {type:\"string\"}"
       ]
     },
     {
@@ -328,10 +468,14 @@
       "source": [
         "from datetime import datetime\n",
         "\n",
-        "TIMESTAMP = datetime.now().strftime(\"%Y%m%d%H%M%S\")\n",
+        "UUID = datetime.now().strftime(\"%Y%m%d%H%M%S\")\n",
         "\n",
-        "if BUCKET_NAME == \"\" or BUCKET_NAME is None or BUCKET_NAME == \"gs://[your-bucket-name]\":\n",
-        "    BUCKET_NAME = \"gs://\" + PROJECT_ID + \"aip-\" + TIMESTAMP"
+        "if (\n",
+        "    BUCKET_NAME == \"\"\n",
+        "    or BUCKET_NAME is None\n",
+        "    or BUCKET_NAME == \"gs://[your-bucket-name-unique]\"\n",
+        "):\n",
+        "    BUCKET_NAME = \"gs://\" + PROJECT_ID + \"aip-\" + UUID"
       ]
     },
     {
@@ -351,7 +495,7 @@
       },
       "outputs": [],
       "source": [
-        "! gsutil mb -l $REGION $BUCKET_NAME"
+        "! gsutil mb -l $REGION -p $PROJECT_ID $BUCKET_NAME"
       ]
     },
     {
@@ -416,10 +560,7 @@
       },
       "outputs": [],
       "source": [
-        "REGION = \"us-central1\"\n",
         "ENDPOINT = \"{}-aiplatform.googleapis.com\".format(REGION)\n",
-        "NETWORK_NAME = \"ucaip-haystack-vpc-network\"  # @param {type:\"string\"}\n",
-        "\n",
         "\n",
         "AUTH_TOKEN = !gcloud auth print-access-token\n",
         "PROJECT_NUMBER = !gcloud projects list --filter=\"PROJECT_ID:'{PROJECT_ID}'\" --format='value(PROJECT_NUMBER)'\n",
@@ -429,10 +570,7 @@
         "\n",
         "print(\"ENDPOINT: {}\".format(ENDPOINT))\n",
         "print(\"PROJECT_ID: {}\".format(PROJECT_ID))\n",
-        "print(\"REGION: {}\".format(REGION))\n",
-        "\n",
-        "!gcloud config set project {PROJECT_ID}\n",
-        "!gcloud config set ai_platform/region {REGION}"
+        "print(\"REGION: {}\".format(REGION))"
       ]
     },
     {
@@ -523,12 +661,13 @@
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "id": "QuVl8DrWG8NS"
       },
       "source": [
-        "Upload the training data to GCS."
+        "Upload the training data to Google Cloud Storage"
       ]
     },
     {
@@ -539,9 +678,9 @@
       },
       "outputs": [],
       "source": [
-        "# NOTE: Everything in this GCS DIR will be DELETED before uploading the data.\n",
+        "# NOTE: Everything in this Google Cloud Storage directory will be DELETED before uploading the data\n",
         "\n",
-        "! gsutil rm -rf {BUCKET_NAME}/*"
+        "! gsutil rm -raf {BUCKET_NAME}/** 2> /dev/null || true"
       ]
     },
     {
@@ -567,21 +706,23 @@
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "id": "mglUPwHpJH98"
       },
       "source": [
-        "## Create Indexes\n"
+        "## Create the indexes\n"
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "id": "qhIBCQ7dDSbW"
       },
       "source": [
-        "### Create ANN Index (for Production Usage)"
+        "### Create Vertex AI Matching Engine index (for production usage)"
       ]
     },
     {
@@ -597,6 +738,16 @@
         ")"
       ]
     },
+    {
+      "attachments": {},
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "14e1ed031d66"
+      },
+      "source": [
+        "Set constants"
+      ]
+    },
     {
       "cell_type": "code",
       "execution_count": null,
@@ -611,14 +762,15 @@
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "id": "svLYiDf0OD2G"
       },
       "source": [
-        "Create the ANN index configuration:\n",
+        "#### Create the Vertex AI Matching Engine index configuration\n",
         "\n",
-        "Please read the documentation to understand the various configuration parameters that can be used to tune the index\n"
+        "Please read the [documentation](https://cloud.google.com/vertex-ai/docs/matching-engine/configuring-indexes) to understand the various configuration parameters that can be used to tune the index"
       ]
     },
     {
@@ -656,9 +808,9 @@
         "    }\n",
         ")\n",
         "\n",
-        "ann_index = {\n",
+        "matching_engine_index = {\n",
         "    \"display_name\": DISPLAY_NAME,\n",
-        "    \"description\": \"Glove 100 ANN index\",\n",
+        "    \"description\": \"Glove 100 Vertex AI Matching Engine Index\",\n",
         "    \"metadata\": struct_pb2.Value(struct_value=metadata),\n",
         "}"
       ]
@@ -671,7 +823,9 @@
       },
       "outputs": [],
       "source": [
-        "ann_index = index_client.create_index(parent=PARENT, index=ann_index)"
+        "matching_engine_index = index_client.create_index(\n",
+        "    parent=PARENT, index=matching_engine_index\n",
+        ")"
       ]
     },
     {
@@ -686,7 +840,7 @@
         "# This will take ~45 min.\n",
         "\n",
         "while True:\n",
-        "    if ann_index.done():\n",
+        "    if matching_engine_index.done():\n",
         "        break\n",
         "    print(\"Poll the operation to create index...\")\n",
         "    time.sleep(60)"
@@ -700,17 +854,18 @@
       },
       "outputs": [],
       "source": [
-        "INDEX_RESOURCE_NAME = ann_index.result().name\n",
+        "INDEX_RESOURCE_NAME = matching_engine_index.result().name\n",
         "INDEX_RESOURCE_NAME"
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "id": "kSsqZuyoA1SG"
       },
       "source": [
-        "### Create Brute Force Index (for Ground Truth)\n",
+        "### Create brute force index (for ground truth)\n",
         "\n",
         "The brute force index uses a naive brute force method to find the nearest neighbors. This method is not fast or efficient. Hence brute force indices are not recommended for production usage. They are to be used to find the \"ground truth\" set of neighbors, so that the \"ground truth\" set can be used to measure recall of the indices being tuned for production usage. To ensure an apples to apples comparison, the `distanceMeasureType` and `featureNormType`, `dimensions` of the brute force index should match those of the production indices being tuned.\n",
         "\n",
@@ -725,8 +880,6 @@
       },
       "outputs": [],
       "source": [
-        "from google.protobuf import *\n",
-        "\n",
         "algorithmConfig = struct_pb2.Struct(\n",
         "    fields={\"bruteForceConfig\": struct_pb2.Value(struct_value=struct_pb2.Struct())}\n",
         ")\n",
@@ -796,12 +949,13 @@
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "id": "mglUPwHpJH98"
       },
       "source": [
-        "## Update Indexes\n",
+        "## Update the indexes\n",
         "\n",
         "Create incremental data file.\n"
       ]
@@ -863,10 +1017,10 @@
         "    }\n",
         ")\n",
         "\n",
-        "ann_index = {\n",
+        "matching_engine_index = {\n",
         "    \"name\": INDEX_RESOURCE_NAME,\n",
         "    \"display_name\": DISPLAY_NAME,\n",
-        "    \"description\": \"Glove 100 ANN index\",\n",
+        "    \"description\": \"Glove 100 Vertex AI Matching Engine Index\",\n",
         "    \"metadata\": struct_pb2.Value(struct_value=metadata),\n",
         "}"
       ]
@@ -879,7 +1033,7 @@
       },
       "outputs": [],
       "source": [
-        "ann_index = index_client.update_index(index=ann_index)"
+        "matching_engine_index = index_client.update_index(index=matching_engine_index)"
       ]
     },
     {
@@ -894,7 +1048,7 @@
         "# This will take ~45 min.\n",
         "\n",
         "while True:\n",
-        "    if ann_index.done():\n",
+        "    if matching_engine_index.done():\n",
         "        break\n",
         "    print(\"Poll the operation to update index...\")\n",
         "    time.sleep(60)"
@@ -908,17 +1062,18 @@
       },
       "outputs": [],
       "source": [
-        "INDEX_RESOURCE_NAME = ann_index.result().name\n",
+        "INDEX_RESOURCE_NAME = matching_engine_index.result().name\n",
         "INDEX_RESOURCE_NAME"
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "id": "qV2xjAnDDObD"
       },
       "source": [
-        "## Create an IndexEndpoint with VPC Network"
+        "## Create an index endpoint with VPC network"
       ]
     },
     {
@@ -997,21 +1152,23 @@
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "id": "np2cgVuuIe9k"
       },
       "source": [
-        "## Deploy Indexes"
+        "## Deploy the indexes"
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "id": "8Ew1UgcIIiJG"
       },
       "source": [
-        "### Deploy ANN Index"
+        "### Deploy a Vertex AI Matching Engine index"
       ]
     },
     {
@@ -1022,7 +1179,7 @@
       },
       "outputs": [],
       "source": [
-        "DEPLOYED_INDEX_ID = \"ann_glove_deployed\""
+        "DEPLOYED_INDEX_ID = \"matching_engine_glove_deployed\""
       ]
     },
     {
@@ -1033,13 +1190,23 @@
       },
       "outputs": [],
       "source": [
-        "deploy_ann_index = {\n",
+        "deploy_matching_engine_index = {\n",
         "    \"id\": DEPLOYED_INDEX_ID,\n",
         "    \"display_name\": DEPLOYED_INDEX_ID,\n",
         "    \"index\": INDEX_RESOURCE_NAME,\n",
         "}"
       ]
     },
+    {
+      "attachments": {},
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "cb6d956d7419"
+      },
+      "source": [
+        "If errors occur with the next command wait some minutes for the index endpoint to be created and retry."
+      ]
+    },
     {
       "cell_type": "code",
       "execution_count": null,
@@ -1049,7 +1216,7 @@
       "outputs": [],
       "source": [
         "r = index_endpoint_client.deploy_index(\n",
-        "    index_endpoint=INDEX_ENDPOINT_NAME, deployed_index=deploy_ann_index\n",
+        "    index_endpoint=INDEX_ENDPOINT_NAME, deployed_index=deploy_matching_engine_index\n",
         ")"
       ]
     },
@@ -1082,12 +1249,13 @@
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "id": "RNZnXmO5AhDO"
       },
       "source": [
-        "### Deploy Brute Force Index"
+        "### Deploy brute force index"
       ]
     },
     {
@@ -1158,12 +1326,13 @@
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "id": "6LCGvBNvBd8D"
       },
       "source": [
-        "## Create Online Queries\n",
+        "## Create online queries\n",
         "\n",
         "After you built your indexes, you may query against the deployed index through the online querying gRPC API (Match service) within the virtual machine instances from the same region (for example 'us-central1' in this tutorial).  \n",
         "\n",
@@ -1178,7 +1347,15 @@
         "\n",
         "* Compile the protocal buffer (see below)\n",
         "* Obtain the index endpoint\n",
-        "* Use a code-generated stub to make the call, passing the parameter values"
+        "* Use a code-generated stub to make the call, passing the parameter values\n",
+        "\n",
+        "### Troubleshooting connectivity issues\n",
+        "\n",
+        "In case you have connectivity errors please perform the following:\n",
+        "\n",
+        "* Verify that the index endpoint, index, and VPC are all in the same Google Cloud project\n",
+        "* Verify that the index endpoint, index, and VPC are all in the same region and it is a valid (e.g. us-central1)\n",
+        "* Verify the Network does not have a firewall rule which denies all egress connections. Else, disable this rule or overwrite it with another rule that allows connection to the index endpoint IP"
       ]
     },
     {
@@ -1351,12 +1528,13 @@
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "id": "8wXTSgz1Bl0x"
       },
       "source": [
-        "Obtain the Private Endpoint: "
+        "Obtain the private endpoint: "
       ]
     },
     {
@@ -1521,12 +1699,13 @@
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "id": "_mNwdU9_B_Ez"
       },
       "source": [
-        "### Batch Query\n",
+        "## Submit a batch query\n",
         "\n",
         "You can run multiple queries in a single RPC call using the BatchMatch API:"
       ]
@@ -1764,18 +1943,20 @@
         "]\n",
         "\n",
         "batch_request = match_service_pb2.BatchMatchRequest()\n",
-        "batch_request_ann = match_service_pb2.BatchMatchRequest.BatchMatchRequestPerIndex()\n",
+        "batch_request_matching_engine = (\n",
+        "    match_service_pb2.BatchMatchRequest.BatchMatchRequestPerIndex()\n",
+        ")\n",
         "batch_request_brute_force = (\n",
         "    match_service_pb2.BatchMatchRequest.BatchMatchRequestPerIndex()\n",
         ")\n",
-        "batch_request_ann.deployed_index_id = DEPLOYED_INDEX_ID\n",
+        "batch_request_matching_engine.deployed_index_id = DEPLOYED_INDEX_ID\n",
         "batch_request_brute_force.deployed_index_id = DEPLOYED_BRUTE_FORCE_INDEX_ID\n",
         "for query in queries:\n",
-        "    batch_request_ann.requests.append(get_request(query, DEPLOYED_INDEX_ID))\n",
+        "    batch_request_matching_engine.requests.append(get_request(query, DEPLOYED_INDEX_ID))\n",
         "    batch_request_brute_force.requests.append(\n",
         "        get_request(query, DEPLOYED_BRUTE_FORCE_INDEX_ID)\n",
         "    )\n",
-        "batch_request.requests.append(batch_request_ann)\n",
+        "batch_request.requests.append(batch_request_matching_engine)\n",
         "batch_request.requests.append(batch_request_brute_force)\n",
         "\n",
         "response = stub.BatchMatch(batch_request)\n",
@@ -1783,14 +1964,15 @@
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "id": "_mNwdU9_B_Ez"
       },
       "source": [
-        "### Compute Recall\n",
+        "### Compute the recall metric\n",
         "\n",
-        "Use deployed brute force Index as the ground truth to calculate the recall of ANN Index:"
+        "Use the deployed brute force index as the ground truth to calculate the recall of the Vertex AI Matching Engine index:"
       ]
     },
     {
@@ -1835,6 +2017,7 @@
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "id": "TpV-iwP9qw9c"
@@ -1844,7 +2027,18 @@
         "\n",
         "To clean up all Google Cloud resources used in this project, you can [delete the Google Cloud\n",
         "project](https://cloud.google.com/resource-manager/docs/creating-managing-projects#shutting_down_projects) you used for the tutorial.\n",
-        "You can also manually delete resources that you created by running the following code."
+        "\n",
+        "Otherwise, you can delete the individual resources you created in this tutorial:"
+      ]
+    },
+    {
+      "attachments": {},
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "390c331dc7d9"
+      },
+      "source": [
+        "### Delete the Vertex AI Matching Engine resources"
       ]
     },
     {
@@ -1869,6 +2063,31 @@
       "source": [
         "index_endpoint_client.delete_index_endpoint(name=INDEX_ENDPOINT_NAME)"
       ]
+    },
+    {
+      "attachments": {},
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "ff14a85c85fb"
+      },
+      "source": [
+        "### Delete the Google Cloud Storage bucket"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "68d4781faac4"
+      },
+      "outputs": [],
+      "source": [
+        "import os\n",
+        "\n",
+        "delete_bucket = False\n",
+        "if delete_bucket or os.getenv(\"IS_TESTING\"):\n",
+        "    ! gsutil -m rm -r $BUCKET_NAME"
+      ]
     }
   ],
   "metadata": {

notebooks/community/ml_ops/stage3/get_started_vertex_pipelines_sklearn_with_prediction.ipynb

@@ -560,6 +560,30 @@
         "    print(\"Service Account:\", SERVICE_ACCOUNT)"
       ]
     },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "set_service_account:pipelines"
+      },
+      "source": [
+        "#### Set service account access for Vertex AI Pipelines\n",
+        "\n",
+        "Run the following commands to grant your service account access to read and write pipeline artifacts in the bucket that you created in the previous step -- you only need to run these once per service account."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "set_service_account:pipelines"
+      },
+      "outputs": [],
+      "source": [
+        "! gsutil iam ch serviceAccount:{SERVICE_ACCOUNT}:roles/storage.objectCreator $BUCKET_URI\n",
+        "\n",
+        "! gsutil iam ch serviceAccount:{SERVICE_ACCOUNT}:roles/storage.objectViewer $BUCKET_URI"
+      ]
+    },
     {
       "cell_type": "markdown",
       "metadata": {

notebooks/community/ml_ops/stage3/get_started_vertex_pipelines_xgboost_with_experiments.ipynb

@@ -557,6 +557,30 @@
         "    print(\"Service Account:\", SERVICE_ACCOUNT)"
       ]
     },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "set_service_account:pipelines"
+      },
+      "source": [
+        "#### Set service account access for Vertex AI Pipelines\n",
+        "\n",
+        "Run the following commands to grant your service account access to read and write pipeline artifacts in the bucket that you created in the previous step -- you only need to run these once per service account."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 13,
+      "metadata": {
+        "id": "set_service_account:pipelines"
+      },
+      "outputs": [],
+      "source": [
+        "! gsutil iam ch serviceAccount:{SERVICE_ACCOUNT}:roles/storage.objectCreator $BUCKET_URI\n",
+        "\n",
+        "! gsutil iam ch serviceAccount:{SERVICE_ACCOUNT}:roles/storage.objectViewer $BUCKET_URI"
+      ]
+    },
     {
       "cell_type": "markdown",
       "metadata": {

notebooks/community/ml_ops/stage3/get_started_vertex_pipelines_xgboost_with_prediction.ipynb

@@ -557,6 +557,30 @@
         "    print(\"Service Account:\", SERVICE_ACCOUNT)"
       ]
     },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "set_service_account:pipelines"
+      },
+      "source": [
+        "#### Set service account access for Vertex AI Pipelines\n",
+        "\n",
+        "Run the following commands to grant your service account access to read and write pipeline artifacts in the bucket that you created in the previous step -- you only need to run these once per service account."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 13,
+      "metadata": {
+        "id": "set_service_account:pipelines"
+      },
+      "outputs": [],
+      "source": [
+        "! gsutil iam ch serviceAccount:{SERVICE_ACCOUNT}:roles/storage.objectCreator $BUCKET_URI\n",
+        "\n",
+        "! gsutil iam ch serviceAccount:{SERVICE_ACCOUNT}:roles/storage.objectViewer $BUCKET_URI"
+      ]
+    },
     {
       "cell_type": "markdown",
       "metadata": {

notebooks/community/ml_ops/stage3/get_started_with_hpt_pipeline_components.ipynb

@@ -1333,7 +1333,6 @@
         "Next, you compile the pipeline and then exeute it. The pipeline takes the following parameters, which are passed as the dictionary `parameter_values`:\n",
         "\n",
         "- `display_name`: A human readable name for the pipeline job.\n",
-        "- `import_file`: The Cloud Storage location to the dataset.\n",
         "- `worker_pool_specs`: The the machine and container, and auto-scaling requirements, as well as command line arguments.\n",
         "- `study_spec_metrics`: The metrics to optimize in the study trials.\n",
         "- `study_spec_parameters`: The parameters to tune."

notebooks/community/ml_ops/stage5/get_started_with_vertex_endpoint_and_shared_vm.ipynb

@@ -159,9 +159,9 @@
         "\n",
         "# Install the packages\n",
         "\n",
-        "! pip3 install --upgrade google-cloud-aiplatform $USER_FLAG -q\n",
-        "! pip3 install --upgrade tensorflow $USER_FLAG -q\n",
-        "! pip3 install --upgrade tensorflow-hub $USER_FLAG -q"
+        "! pip3 install --upgrade google-cloud-aiplatform \\\n",
+        "                         tensorflow \\\n",
+        "                         tensorflow-hub $USER_FLAG -q"
       ]
     },
     {
@@ -307,22 +307,29 @@
         "id": "timestamp"
       },
       "source": [
-        "#### Timestamp\n",
+        "#### UUID\n",
         "\n",
-        "If you are in a live tutorial session, you might be using a shared test account or project. To avoid name collisions between users on resources created, you create a timestamp for each instance session, and append the timestamp onto the name of resources you create in this tutorial."
+        "If you are in a live tutorial session, you might be using a shared test account or project. To avoid name collisions between users on resources created, you create a uuid for each instance session, and append it onto the name of resources you create in this tutorial."
       ]
     },
     {
       "cell_type": "code",
       "execution_count": null,
       "metadata": {
-        "id": "timestamp"
+        "id": "84Vdv7R-QEH6"
       },
       "outputs": [],
       "source": [
-        "from datetime import datetime\n",
+        "import random\n",
+        "import string\n",
         "\n",
-        "TIMESTAMP = datetime.now().strftime(\"%Y%m%d%H%M%S\")"
+        "\n",
+        "# Generate a uuid of a specifed length(default=8)\n",
+        "def generate_uuid(length: int = 8) -> str:\n",
+        "    return \"\".join(random.choices(string.ascii_lowercase + string.digits, k=length))\n",
+        "\n",
+        "\n",
+        "UUID = generate_uuid()"
       ]
     },
     {
@@ -421,7 +428,7 @@
       "outputs": [],
       "source": [
         "if BUCKET_URI == \"\" or BUCKET_URI is None or BUCKET_URI == \"gs://[your-bucket-name]\":\n",
-        "    BUCKET_NAME = PROJECT_ID + \"aip-\" + TIMESTAMP\n",
+        "    BUCKET_NAME = PROJECT_ID + \"aip-\" + UUID\n",
         "    BUCKET_URI = \"gs://\" + BUCKET_NAME"
       ]
     },
@@ -523,7 +530,7 @@
         "\n",
         "Setup up the following constants for Vertex AI:\n",
         "\n",
-        "- `API_ENDPOINT`: The Vertex AI API service endpoint for `Endpoint` services."
+        "- `API_ENDPOINT`: The Vertex AI API service endpoint."
       ]
     },
     {
@@ -538,46 +545,10 @@
         "API_ENDPOINT = \"{}-aiplatform.googleapis.com\".format(REGION)\n",
         "\n",
         "# Vertex location root path for your dataset, model and endpoint resources\n",
-        "PARENT = \"projects/\" + PROJECT_ID + \"/locations/\" + REGION"
-      ]
-    },
-    {
-      "cell_type": "markdown",
-      "metadata": {
-        "id": "clients:metadata"
-      },
-      "source": [
-        "## Set up clients\n",
+        "PARENT = \"projects/\" + PROJECT_ID + \"/locations/\" + REGION\n",
         "\n",
-        "The Vertex  works as a client/server model. On your side (the Python script) you will create a client that sends requests and receives responses from the Vertex AI server.\n",
-        "\n",
-        "You will use different clients in this tutorial for different steps in the workflow. So set them all up upfront.\n",
-        "\n",
-        "- Endpoint Service for creating endpoints, and deploying models to endpoints."
-      ]
-    },
-    {
-      "cell_type": "code",
-      "execution_count": null,
-      "metadata": {
-        "id": "clients:metadata"
-      },
-      "outputs": [],
-      "source": [
         "# client options same for all services\n",
-        "client_options = {\"api_endpoint\": API_ENDPOINT}\n",
-        "\n",
-        "\n",
-        "def create_endpoint_client():\n",
-        "    client = aip_beta.EndpointServiceClient(client_options=client_options)\n",
-        "    return client\n",
-        "\n",
-        "\n",
-        "clients = {}\n",
-        "clients[\"endpoint\"] = create_endpoint_client()\n",
-        "\n",
-        "for client in clients.items():\n",
-        "    print(client)"
+        "client_options = {\"api_endpoint\": API_ENDPOINT}"
       ]
     },
     {
@@ -592,7 +563,7 @@
         "\n",
         "Set the variables `DEPLOY_GPU/DEPLOY_NGPU` to use a container image supporting a GPU and the number of GPUs allocated to the virtual machine (VM) instance. For example, to use a GPU container image with 4 Nvidia Telsa K80 GPUs allocated to each VM, you would specify:\n",
         "\n",
-        "    (aip.AcceleratorType.NVIDIA_TESLA_K80, 4)\n",
+        "    (aip.gapic.AcceleratorType.NVIDIA_TESLA_K80, 4)\n",
         "\n",
         "\n",
         "Otherwise specify `(None, None)` to use a container image to run on a CPU.\n",
@@ -902,7 +873,7 @@
       "outputs": [],
       "source": [
         "model_icn = aiplatform.Model.upload(\n",
-        "    display_name=\"icn_\" + TIMESTAMP,\n",
+        "    display_name=\"icn_\" + UUID,\n",
         "    artifact_uri=MODEL_ICN_DIR,\n",
         "    serving_container_image_uri=DEPLOY_IMAGE,\n",
         ")\n",
@@ -1013,7 +984,7 @@
       "outputs": [],
       "source": [
         "model_use = aiplatform.Model.upload(\n",
-        "    display_name=\"icn_\" + TIMESTAMP,\n",
+        "    display_name=\"icn_\" + UUID,\n",
         "    artifact_uri=MODEL_USE_DIR,\n",
         "    serving_container_image_uri=DEPLOY_IMAGE,\n",
         ")\n",
@@ -1029,64 +1000,55 @@
       "source": [
         "## Creating a deployment resource pool\n",
         "\n",
-        "Currently, creating deploynent resource pools is only supported via the REST-based API (e.g., CURL).\n",
+        "Currently, creating deploynent resource pools is only supported via the REST-based API (e.g., CURL) and GAPIC APIs (Python).\n",
         "\n",
-        "Use `CreateDeploymentResourcePool` API to create a resource pool, with the following configuration:\n",
+        "Use `create_deployment_resource_pool` API to create a resource pool, with the following configuration:\n",
         "\n",
         "- `dedicated_resources`: Compute (HW) resources to allocate for the shared vm.\n",
         "- `min_replica_count`: Auto-scaling, the minimum number of compute nodes.\n",
         "- `max_replica_count`: Auto-scaling, the maximum number of compute nodes.\n",
         "\n",
-        "Learn more about [Deployment Resource Pools]()."
+        "Learn more about [Deployment Resource Pools](https://googleapis.dev/python/aiplatform/latest/aiplatform_v1beta1/deployment_resource_pool_service.html)."
       ]
     },
     {
       "cell_type": "code",
       "execution_count": null,
       "metadata": {
-        "id": "YiBmoiWYcMQt"
+        "id": "90c51b6cf34a"
       },
       "outputs": [],
       "source": [
-        "DEPLOYMENT_RESOURCE_POOL_ID = \"shared-vm\"  # @param {type: \"string\"}"
-      ]
-    },
-    {
-      "cell_type": "code",
-      "execution_count": null,
-      "metadata": {
-        "id": "0CHPJ4h-Slgs"
-      },
-      "outputs": [],
-      "source": [
-        "import json\n",
-        "import pprint\n",
-        "pp = pprint.PrettyPrinter(indent=4)\n",
-        "\n",
+        "DEPLOYMENT_RESOURCE_POOL_ID = f\"shared-vm-{UUID}\"  # @param {type: \"string\"}\n",
         "MIN_NODES = 1\n",
         "MAX_NODES = 2\n",
         "\n",
-        "CREATE_RP_PAYLOAD = {\n",
-        "  \"deployment_resource_pool\":{\n",
-        "    \"dedicated_resources\":{\n",
-        "      \"machine_spec\":{\n",
-        "        \"machine_type\": DEPLOY_COMPUTE\n",
-        "      },\n",
-        "      \"min_replica_count\": MIN_NODES, \n",
-        "      \"max_replica_count\": MAX_NODES\n",
-        "    }\n",
-        "  },\n",
-        "  \"deployment_resource_pool_id\":DEPLOYMENT_RESOURCE_POOL_ID\n",
-        "}\n",
-        "CREATE_RP_REQUEST=json.dumps(CREATE_RP_PAYLOAD)\n",
-        "pp.pprint(\"CREATE_RP_REQUEST: \" + CREATE_RP_REQUEST)\n",
+        "# Initialize request argument(s)\n",
+        "deployment_resource_pool = aip_beta.DeploymentResourcePool()\n",
+        "deployment_resource_pool.dedicated_resources.min_replica_count = MIN_NODES\n",
+        "deployment_resource_pool.dedicated_resources.max_replica_count = MAX_NODES\n",
+        "deployment_resource_pool.dedicated_resources.machine_spec.machine_type = DEPLOY_COMPUTE\n",
+        "if DEPLOY_NGPU:\n",
+        "    deployment_resource_pool.dedicated_resources.machine_spec.accelerator_type = DEPLOY_GPU\n",
+        "    deployment_resource_pool.dedicated_resources.machine_spec.accelerator_count = DEPLOY_NGPU\n",
         "\n",
-        "! curl \\\n",
-        "-X POST \\\n",
-        "-H \"Authorization: Bearer $(gcloud auth print-access-token)\" \\\n",
-        "-H \"Content-Type: application/json\" \\\n",
-        "https://{REGION}-aiplatform.googleapis.com/v1beta1/projects/{PROJECT_ID}/locations/{REGION}/deploymentResourcePools \\\n",
-        "-d '{CREATE_RP_REQUEST}'"
+        "request = aip_beta.CreateDeploymentResourcePoolRequest(\n",
+        "    parent=f\"projects/{PROJECT_ID}/locations/{REGION}\",\n",
+        "    deployment_resource_pool=deployment_resource_pool,\n",
+        "    deployment_resource_pool_id=DEPLOYMENT_RESOURCE_POOL_ID,\n",
+        ")\n",
+        "\n",
+        "pool_client = aip_beta.services.deployment_resource_pool_service.DeploymentResourcePoolServiceClient(\n",
+        "    client_options=client_options\n",
+        ")\n",
+        "\n",
+        "op = pool_client.create_deployment_resource_pool(request=request)\n",
+        "print(op)\n",
+        "\n",
+        "result = op.result()\n",
+        "print(result)\n",
+        "\n",
+        "deployment_pool_id = result.name"
       ]
     },
     {
@@ -1099,21 +1061,19 @@
         "\n",
         "Use `GetDeploymentResourcePool` API to check out the deploynent resource pool that you created. \n",
         "\n",
-        "Learn more about [Get Deployment Resource Pool](https://source.corp.google.com/piper///depot/google3/google/cloud/aiplatform/master/deployment_resource_pool_service.proto;l=75?q=deployment_resource_pool&sq=package:piper%20file:%2F%2Fdepot%2Fgoogle3%20-file:google3%2Fexperimental)."
+        "Learn more about [Get Deployment Resource Pool](https://googleapis.dev/python/aiplatform/latest/aiplatform_v1beta1/deployment_resource_pool_service.html)."
       ]
     },
     {
       "cell_type": "code",
       "execution_count": null,
       "metadata": {
-        "id": "6wTLyhPraFah"
+        "id": "b740253903c0"
       },
       "outputs": [],
       "source": [
-        "! curl -X GET \\\n",
-        "-H \"Authorization: Bearer $(gcloud auth print-access-token)\" \\\n",
-        "-H \"Content-Type: application/json\" \\\n",
-        "https://{REGION}-aiplatform.googleapis.com/v1beta1/projects/{PROJECT_ID}/locations/{REGION}/deploymentResourcePools/{DEPLOYMENT_RESOURCE_POOL_ID}"
+        "response = pool_client.get_deployment_resource_pool(name=deployment_pool_id)\n",
+        "print(response)"
       ]
     },
     {
@@ -1126,21 +1086,22 @@
         "\n",
         "Use `ListDeploymentResourcePools` API to list all the deployment resource pools. \n",
         "\n",
-        "Learn more about [Listing Deployment Resource Pools](https://source.corp.google.com/piper///depot/google3/google/cloud/aiplatform/master/deployment_resource_pool_service.proto;l=101?q=deployment_resource_pool&sq=package:piper%20file:%2F%2Fdepot%2Fgoogle3%20-file:google3%2Fexperimental)."
+        "Learn more about [Listing Deployment Resource Pools](https://googleapis.dev/python/aiplatform/latest/aiplatform_v1beta1/deployment_resource_pool_service.html)."
       ]
     },
     {
       "cell_type": "code",
       "execution_count": null,
       "metadata": {
-        "id": "Pxls4sNnaltU"
+        "id": "3ebfd007bff2"
       },
       "outputs": [],
       "source": [
-        "! curl -X GET \\\n",
-        "-H \"Authorization: Bearer $(gcloud auth print-access-token)\" \\\n",
-        "-H \"Content-Type: application/json\" \\\n",
-        "https://{REGION}-aiplatform.googleapis.com/v1beta1/projects/{PROJECT_ID}/locations/{REGION}/deploymentResourcePools"
+        "pools = pool_client.list_deployment_resource_pools(\n",
+        "    parent=f\"projects/{PROJECT_ID}/locations/{REGION}\"\n",
+        ")\n",
+        "for pool in pools:\n",
+        "    print(pool)"
       ]
     },
     {
@@ -1170,11 +1131,11 @@
       },
       "outputs": [],
       "source": [
-        "endpoint_icn = aiplatform.Endpoint.create(display_name=\"icn_\" + TIMESTAMP)\n",
+        "endpoint_icn = aiplatform.Endpoint.create(display_name=\"icn_\" + UUID)\n",
         "\n",
         "print(endpoint_icn)\n",
         "\n",
-        "endpoint_use = aiplatform.Endpoint.create(display_name=\"use_\" + TIMESTAMP)\n",
+        "endpoint_use = aiplatform.Endpoint.create(display_name=\"use_\" + UUID)\n",
         "\n",
         "print(endpoint_use)"
       ]
@@ -1204,6 +1165,12 @@
       },
       "outputs": [],
       "source": [
+        "import json\n",
+        "import pprint\n",
+        "\n",
+        "pp = pprint.PrettyPrinter(indent=4)\n",
+        "\n",
+        "\n",
         "SHARED_RESOURCE = \"projects/{project_id}/locations/{region}/deploymentResourcePools/{deployment_resource_pool_id}\".format(\n",
         "    project_id=PROJECT_ID,\n",
         "    region=REGION,\n",
@@ -1363,18 +1330,27 @@
         "    time.sleep(30)"
       ]
     },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "52248c450776"
+      },
+      "source": [
+        "### Get deployment details for the endpoint\n",
+        "\n",
+        "List the deployed models on the endpoint."
+      ]
+    },
     {
       "cell_type": "code",
       "execution_count": null,
       "metadata": {
-        "id": "86a659bf60f0"
+        "id": "3b768614e7c6"
       },
       "outputs": [],
       "source": [
-        "! curl -X GET \\\n",
-        " -H \"Authorization: Bearer $(gcloud auth print-access-token)\" \\\n",
-        " -H \"Content-Type: application/json\" \\\n",
-        "https://{REGION}-aiplatform.googleapis.com/v1/projects/759209241365/locations/us-central1/endpoints/2259566763823857664"
+        "print(endpoint_icn.list_models())\n",
+        "print(endpoint_use.list_models())"
       ]
     },
     {
@@ -1557,21 +1533,19 @@
       "source": [
         "#### Delete the `DeploymentResourcePool`\n",
         "\n",
-        "The method 'delete()' will delete your deployment resource pool."
+        "The method 'delete_deployment_resource_pool()' will delete your deployment resource pool."
       ]
     },
     {
       "cell_type": "code",
       "execution_count": null,
       "metadata": {
-        "id": "ac40cc1d594a"
+        "id": "b76a4de1e57e"
       },
       "outputs": [],
       "source": [
-        "! curl -X DELETE \\\n",
-        "-H \"Authorization: Bearer $(gcloud auth print-access-token)\" \\\n",
-        "-H \"Content-Type: application/json\" \\\n",
-        "https://{REGION}-aiplatform.googleapis.com/v1beta1/projects/{PROJECT_ID}/locations/{REGION}/deploymentResourcePools/{DEPLOYMENT_RESOURCE_POOL_ID}"
+        "response = pool_client.delete_deployment_resource_pool(name=deployment_pool_id)\n",
+        "print(response)"
       ]
     },
     {

notebooks/community/model_garden/model_garden_pytorch_blip_image_captioning.ipynb

@@ -0,0 +1,390 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "7d9bbf86da5e"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2023 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "2bd716bf3e39"
+      },
+      "source": [
+        "# Vertex AI Model Garden - BLIP Image Captioning\n",
+        "\n",
+        "<table align=\"left\">\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_blip_image_captioning.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_blip_image_captioning.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>                                                                                               <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/notebooks/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_blip_image_captioning.ipynb\">\n",
+        "      <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d8cd12648da4"
+      },
+      "source": [
+        "## Overview\n",
+        "\n",
+        "This notebook demonstrates deploying the pre-trained [BLIP Image Captioning](https://huggingface.co/Salesforce/blip-image-captioning-base) model on Vertex AI for online prediction.\n",
+        "\n",
+        "### Objective\n",
+        "\n",
+        "- Upload the model to [Model Registry](https://cloud.google.com/vertex-ai/docs/model-registry/introduction).\n",
+        "- Deploy the model on [Endpoint](https://cloud.google.com/vertex-ai/docs/predictions/using-private-endpoints).\n",
+        "- Run online predictions for image captioning.\n",
+        "\n",
+        "### Costs\n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI\n",
+        "* Cloud Storage\n",
+        "\n",
+        "Learn about [Vertex AI pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage pricing](https://cloud.google.com/storage/pricing), and use the [Pricing Calculator](https://cloud.google.com/products/calculator/) to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "264c07757582"
+      },
+      "source": [
+        "## Setup environment\n",
+        "\n",
+        "**NOTE**: Jupyter runs lines prefixed with `!` as shell commands, and it interpolates Python variables prefixed with `$` into these commands."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d73ffa0c0b83"
+      },
+      "source": [
+        "### Colab only"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "2707b02ef5df"
+      },
+      "outputs": [],
+      "source": [
+        "!pip3 install --upgrade google-cloud-aiplatform"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b60a4d7100bf"
+      },
+      "outputs": [],
+      "source": [
+        "from google.colab import auth as google_auth\n",
+        "\n",
+        "google_auth.authenticate_user()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0f826ff482a2"
+      },
+      "source": [
+        "### Setup Google Cloud project\n",
+        "\n",
+        "1. [Select or create a Google Cloud project](https://console.cloud.google.com/cloud-resource-manager). When you first create an account, you get a $300 free credit towards your compute/storage costs.\n",
+        "\n",
+        "1. [Make sure that billing is enabled for your project](https://cloud.google.com/billing/docs/how-to/modify-project).\n",
+        "\n",
+        "1. [Enable the Vertex AI API and Compute Engine API](https://console.cloud.google.com/flows/enableapi?apiid=aiplatform.googleapis.com,compute_component).\n",
+        "\n",
+        "1. [Create a Cloud Storage bucket](https://cloud.google.com/storage/docs/creating-buckets) for storing experiment outputs."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "8958ebc71868"
+      },
+      "source": [
+        "Fill following variables for experiments environment:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "9db30f827a65"
+      },
+      "outputs": [],
+      "source": [
+        "# Cloud project id.\n",
+        "PROJECT_ID = \"\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# The region you want to launch jobs in.\n",
+        "REGION = \"us-central1\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# The Cloud Storage bucket for storing experiments output. Fill it without the 'gs://' prefix.\n",
+        "GCS_BUCKET = \"\"  # @param {type:\"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "92f16e22c20b"
+      },
+      "source": [
+        "Initialize Vertex AI API:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "1680c257acfb"
+      },
+      "outputs": [],
+      "source": [
+        "from google.cloud import aiplatform\n",
+        "\n",
+        "aiplatform.init(project=PROJECT_ID, location=REGION, staging_bucket=GCS_BUCKET)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "6ca48b699d17"
+      },
+      "source": [
+        "### Define constants"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "de9882ea89ea"
+      },
+      "outputs": [],
+      "source": [
+        "# The pre-built serving docker image. It contains serving scripts and models.\n",
+        "SERVE_DOCKER_URI = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/pytorch-transformers-serve:latest\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "10188266a5cd"
+      },
+      "source": [
+        "### Define common functions"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "cac4478ae098"
+      },
+      "outputs": [],
+      "source": [
+        "import base64\n",
+        "import os\n",
+        "from datetime import datetime\n",
+        "from io import BytesIO\n",
+        "\n",
+        "import requests\n",
+        "from google.cloud import aiplatform\n",
+        "from PIL import Image\n",
+        "\n",
+        "\n",
+        "def create_job_name(prefix):\n",
+        "    user = os.environ.get(\"USER\")\n",
+        "    now = datetime.now().strftime(\"%Y%m%d_%H%M%S\")\n",
+        "    job_name = f\"{prefix}-{user}-{now}\"\n",
+        "    return job_name\n",
+        "\n",
+        "\n",
+        "def download_image(url):\n",
+        "    response = requests.get(url)\n",
+        "    return Image.open(BytesIO(response.content))\n",
+        "\n",
+        "\n",
+        "def image_to_base64(image, format=\"JPEG\"):\n",
+        "    buffer = BytesIO()\n",
+        "    image.save(buffer, format=format)\n",
+        "    image_str = base64.b64encode(buffer.getvalue()).decode(\"utf-8\")\n",
+        "    return image_str\n",
+        "\n",
+        "\n",
+        "def base64_to_image(image_str):\n",
+        "    image = Image.open(BytesIO(base64.b64decode(image_str)))\n",
+        "    return image\n",
+        "\n",
+        "\n",
+        "def image_grid(imgs, rows=2, cols=2):\n",
+        "    w, h = imgs[0].size\n",
+        "    grid = Image.new(\"RGB\", size=(cols * w, rows * h))\n",
+        "    for i, img in enumerate(imgs):\n",
+        "        grid.paste(img, box=(i % cols * w, i // cols * h))\n",
+        "    return grid\n",
+        "\n",
+        "\n",
+        "def deploy_model(model_id, task):\n",
+        "    model_name = \"blip-image-captioning\"\n",
+        "    endpoint = aiplatform.Endpoint.create(display_name=f\"{model_name}-endpoint\")\n",
+        "    serving_env = {\n",
+        "        \"MODEL_ID\": model_id,\n",
+        "        \"TASK\": task,\n",
+        "    }\n",
+        "    # If the model_id is a GCS path, use artifact_uri to pass it to serving docker.\n",
+        "    artifact_uri = model_id if model_id.startswith(\"gs://\") else None\n",
+        "    model = aiplatform.Model.upload(\n",
+        "        display_name=model_name,\n",
+        "        serving_container_image_uri=SERVE_DOCKER_URI,\n",
+        "        serving_container_ports=[7080],\n",
+        "        serving_container_predict_route=\"/predictions/diffusers_serving\",\n",
+        "        serving_container_health_route=\"/ping\",\n",
+        "        serving_container_environment_variables=serving_env,\n",
+        "        artifact_uri=artifact_uri,\n",
+        "    )\n",
+        "    model.deploy(\n",
+        "        endpoint=endpoint,\n",
+        "        machine_type=\"n1-standard-8\",\n",
+        "        accelerator_type=\"NVIDIA_TESLA_T4\",\n",
+        "        accelerator_count=1,\n",
+        "        deploy_request_timeout=1800,\n",
+        "    )\n",
+        "    return model, endpoint"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d2d72ecdb8c9"
+      },
+      "source": [
+        "## Upload and deploy models"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "9448c5f545fa"
+      },
+      "source": [
+        "This section uploads the pre-trained model to Model Registry and deploys it on the Endpoint with 1 T4 GPU.\n",
+        "\n",
+        "The model deployment step will take ~15 minutes to complete.\n",
+        "\n",
+        "Once deployed, you can send images to get descriptions."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b4b46c28d8b1"
+      },
+      "outputs": [],
+      "source": [
+        "model, endpoint = deploy_model(\n",
+        "    model_id=\"Salesforce/blip-image-captioning-base\", task=\"image-to-text\"\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "6be655247cb1"
+      },
+      "outputs": [],
+      "source": [
+        "image = download_image(\"http://images.cocodataset.org/val2017/000000039769.jpg\")\n",
+        "display(image)\n",
+        "\n",
+        "instances = [\n",
+        "    {\"image\": image_to_base64(image)},\n",
+        "]\n",
+        "preds = endpoint.predict(instances=instances).predictions\n",
+        "print(preds)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "db7ffebdb4be"
+      },
+      "source": [
+        "### Clean up resources"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "2ccf3714dbe9"
+      },
+      "outputs": [],
+      "source": [
+        "# Delete models.\n",
+        "model.delete()\n",
+        "\n",
+        "# Undeploy model and delete endpoint.\n",
+        "endpoint.delete(force=True)"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "name": "model_garden_pytorch_blip_image_captioning.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/community/model_garden/model_garden_pytorch_blip_vqa.ipynb

@@ -0,0 +1,392 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "7d9bbf86da5e"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2023 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "2bd716bf3e39"
+      },
+      "source": [
+        "# Vertex AI Model Garden - BLIP VQA\n",
+        "\n",
+        "<table align=\"left\">\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_blip_vqa.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_blip_vqa.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>                                                                                               <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/notebooks/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_blip_vqa.ipynb\">\n",
+        "      <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d8cd12648da4"
+      },
+      "source": [
+        "## Overview\n",
+        "\n",
+        "This notebook demonstrates deploying the pre-trained [BLIP VQA](https://huggingface.co/Salesforce/blip-vqa-base) model on Vertex AI for online prediction.\n",
+        "\n",
+        "### Objective\n",
+        "\n",
+        "- Upload the model to [Model Registry](https://cloud.google.com/vertex-ai/docs/model-registry/introduction).\n",
+        "- Deploy the model on [Endpoint](https://cloud.google.com/vertex-ai/docs/predictions/using-private-endpoints).\n",
+        "- Run online predictions for image captioning.\n",
+        "\n",
+        "### Costs\n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI\n",
+        "* Cloud Storage\n",
+        "\n",
+        "Learn about [Vertex AI pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage pricing](https://cloud.google.com/storage/pricing), and use the [Pricing Calculator](https://cloud.google.com/products/calculator/) to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "264c07757582"
+      },
+      "source": [
+        "## Setup environment\n",
+        "\n",
+        "**NOTE**: Jupyter runs lines prefixed with `!` as shell commands, and it interpolates Python variables prefixed with `$` into these commands."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d73ffa0c0b83"
+      },
+      "source": [
+        "### Colab only"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "2707b02ef5df"
+      },
+      "outputs": [],
+      "source": [
+        "!pip3 install --upgrade google-cloud-aiplatform"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b60a4d7100bf"
+      },
+      "outputs": [],
+      "source": [
+        "from google.colab import auth as google_auth\n",
+        "\n",
+        "google_auth.authenticate_user()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0f826ff482a2"
+      },
+      "source": [
+        "### Setup Google Cloud project\n",
+        "\n",
+        "1. [Select or create a Google Cloud project](https://console.cloud.google.com/cloud-resource-manager). When you first create an account, you get a $300 free credit towards your compute/storage costs.\n",
+        "\n",
+        "1. [Make sure that billing is enabled for your project](https://cloud.google.com/billing/docs/how-to/modify-project).\n",
+        "\n",
+        "1. [Enable the Vertex AI API and Compute Engine API](https://console.cloud.google.com/flows/enableapi?apiid=aiplatform.googleapis.com,compute_component).\n",
+        "\n",
+        "1. [Create a Cloud Storage bucket](https://cloud.google.com/storage/docs/creating-buckets) for storing experiment outputs."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "8958ebc71868"
+      },
+      "source": [
+        "Fill following variables for experiments environment:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "9db30f827a65"
+      },
+      "outputs": [],
+      "source": [
+        "# Cloud project id.\n",
+        "PROJECT_ID = \"\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# The region you want to launch jobs in.\n",
+        "REGION = \"us-central1\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# The Cloud Storage bucket for storing experiments output. Fill it without the 'gs://' prefix.\n",
+        "GCS_BUCKET = \"\"  # @param {type:\"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "92f16e22c20b"
+      },
+      "source": [
+        "Initialize Vertex AI API:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "1680c257acfb"
+      },
+      "outputs": [],
+      "source": [
+        "from google.cloud import aiplatform\n",
+        "\n",
+        "aiplatform.init(project=PROJECT_ID, location=REGION, staging_bucket=GCS_BUCKET)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "6ca48b699d17"
+      },
+      "source": [
+        "### Define constants"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "de9882ea89ea"
+      },
+      "outputs": [],
+      "source": [
+        "# The pre-built serving docker image. It contains serving scripts and models.\n",
+        "SERVE_DOCKER_URI = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/pytorch-transformers-serve:latest\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "10188266a5cd"
+      },
+      "source": [
+        "### Define common functions"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "cac4478ae098"
+      },
+      "outputs": [],
+      "source": [
+        "import base64\n",
+        "import os\n",
+        "from datetime import datetime\n",
+        "from io import BytesIO\n",
+        "\n",
+        "import requests\n",
+        "from google.cloud import aiplatform\n",
+        "from PIL import Image\n",
+        "\n",
+        "\n",
+        "def create_job_name(prefix):\n",
+        "    user = os.environ.get(\"USER\")\n",
+        "    now = datetime.now().strftime(\"%Y%m%d_%H%M%S\")\n",
+        "    job_name = f\"{prefix}-{user}-{now}\"\n",
+        "    return job_name\n",
+        "\n",
+        "\n",
+        "def download_image(url):\n",
+        "    response = requests.get(url)\n",
+        "    return Image.open(BytesIO(response.content))\n",
+        "\n",
+        "\n",
+        "def image_to_base64(image, format=\"JPEG\"):\n",
+        "    buffer = BytesIO()\n",
+        "    image.save(buffer, format=format)\n",
+        "    image_str = base64.b64encode(buffer.getvalue()).decode(\"utf-8\")\n",
+        "    return image_str\n",
+        "\n",
+        "\n",
+        "def base64_to_image(image_str):\n",
+        "    image = Image.open(BytesIO(base64.b64decode(image_str)))\n",
+        "    return image\n",
+        "\n",
+        "\n",
+        "def image_grid(imgs, rows=2, cols=2):\n",
+        "    w, h = imgs[0].size\n",
+        "    grid = Image.new(\"RGB\", size=(cols * w, rows * h))\n",
+        "    for i, img in enumerate(imgs):\n",
+        "        grid.paste(img, box=(i % cols * w, i // cols * h))\n",
+        "    return grid\n",
+        "\n",
+        "\n",
+        "def deploy_model(model_id, task):\n",
+        "    model_name = \"blip-vqa\"\n",
+        "    endpoint = aiplatform.Endpoint.create(display_name=f\"{model_name}-endpoint\")\n",
+        "    serving_env = {\n",
+        "        \"MODEL_ID\": model_id,\n",
+        "        \"TASK\": task,\n",
+        "    }\n",
+        "    # If the model_id is a GCS path, use artifact_uri to pass it to serving docker.\n",
+        "    artifact_uri = model_id if model_id.startswith(\"gs://\") else None\n",
+        "    model = aiplatform.Model.upload(\n",
+        "        display_name=model_name,\n",
+        "        serving_container_image_uri=SERVE_DOCKER_URI,\n",
+        "        serving_container_ports=[7080],\n",
+        "        serving_container_predict_route=\"/predictions/diffusers_serving\",\n",
+        "        serving_container_health_route=\"/ping\",\n",
+        "        serving_container_environment_variables=serving_env,\n",
+        "        artifact_uri=artifact_uri,\n",
+        "    )\n",
+        "    model.deploy(\n",
+        "        endpoint=endpoint,\n",
+        "        machine_type=\"n1-standard-8\",\n",
+        "        accelerator_type=\"NVIDIA_TESLA_T4\",\n",
+        "        accelerator_count=1,\n",
+        "        deploy_request_timeout=1800,\n",
+        "    )\n",
+        "    return model, endpoint"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d2d72ecdb8c9"
+      },
+      "source": [
+        "## Upload and deploy models"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "9448c5f545fa"
+      },
+      "source": [
+        "This section uploads the pre-trained model to Model Registry and deploys it on the Endpoint with 1 T4 GPU.\n",
+        "\n",
+        "The model deployment step will take ~15 minutes to complete.\n",
+        "\n",
+        "Once deployed, you can send images and questions to get answers."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b4b46c28d8b1"
+      },
+      "outputs": [],
+      "source": [
+        "model, endpoint = deploy_model(\n",
+        "    model_id=\"Salesforce/blip-vqa-base\", task=\"visual-question-answering\"\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "6be655247cb1"
+      },
+      "outputs": [],
+      "source": [
+        "image = download_image(\"http://images.cocodataset.org/val2017/000000039769.jpg\")\n",
+        "display(image)\n",
+        "\n",
+        "question = \"Which cat is bigger?\"\n",
+        "instances = [\n",
+        "    {\"image\": image_to_base64(image), \"text\": question},\n",
+        "]\n",
+        "preds = endpoint.predict(instances=instances).predictions\n",
+        "print(question)\n",
+        "print(preds)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "db7ffebdb4be"
+      },
+      "source": [
+        "### Clean up resources"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "2ccf3714dbe9"
+      },
+      "outputs": [],
+      "source": [
+        "# Delete models.\n",
+        "model.delete()\n",
+        "\n",
+        "# Undeploy model and delete endpoint.\n",
+        "endpoint.delete(force=True)"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "name": "model_garden_pytorch_blip_vqa.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/community/model_garden/model_garden_pytorch_clip.ipynb

@@ -0,0 +1,393 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "7d9bbf86da5e"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2023 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "2bd716bf3e39"
+      },
+      "source": [
+        "# Vertex AI Model Garden - CLIP\n",
+        "\n",
+        "<table align=\"left\">\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_clip.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_clip.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>                                                                                               <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/notebooks/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_clip.ipynb\">\n",
+        "      <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d8cd12648da4"
+      },
+      "source": [
+        "## Overview\n",
+        "\n",
+        "This notebook demonstrates deploying the pre-trained [CLIP](https://huggingface.co/openai/clip-vit-base-patch32) model on Vertex AI for online prediction.\n",
+        "\n",
+        "### Objective\n",
+        "\n",
+        "- Upload the model to [Model Registry](https://cloud.google.com/vertex-ai/docs/model-registry/introduction).\n",
+        "- Deploy the model on [Endpoint](https://cloud.google.com/vertex-ai/docs/predictions/using-private-endpoints).\n",
+        "- Run online predictions for image captioning.\n",
+        "\n",
+        "### Costs\n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI\n",
+        "* Cloud Storage\n",
+        "\n",
+        "Learn about [Vertex AI pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage pricing](https://cloud.google.com/storage/pricing), and use the [Pricing Calculator](https://cloud.google.com/products/calculator/) to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "264c07757582"
+      },
+      "source": [
+        "## Setup environment\n",
+        "\n",
+        "**NOTE**: Jupyter runs lines prefixed with `!` as shell commands, and it interpolates Python variables prefixed with `$` into these commands."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d73ffa0c0b83"
+      },
+      "source": [
+        "### Colab only"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "2707b02ef5df"
+      },
+      "outputs": [],
+      "source": [
+        "!pip3 install --upgrade google-cloud-aiplatform"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b60a4d7100bf"
+      },
+      "outputs": [],
+      "source": [
+        "from google.colab import auth as google_auth\n",
+        "\n",
+        "google_auth.authenticate_user()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0f826ff482a2"
+      },
+      "source": [
+        "### Setup Google Cloud project\n",
+        "\n",
+        "1. [Select or create a Google Cloud project](https://console.cloud.google.com/cloud-resource-manager). When you first create an account, you get a $300 free credit towards your compute/storage costs.\n",
+        "\n",
+        "1. [Make sure that billing is enabled for your project](https://cloud.google.com/billing/docs/how-to/modify-project).\n",
+        "\n",
+        "1. [Enable the Vertex AI API and Compute Engine API](https://console.cloud.google.com/flows/enableapi?apiid=aiplatform.googleapis.com,compute_component).\n",
+        "\n",
+        "1. [Create a Cloud Storage bucket](https://cloud.google.com/storage/docs/creating-buckets) for storing experiment outputs."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "8958ebc71868"
+      },
+      "source": [
+        "Fill following variables for experiments environment:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "9db30f827a65"
+      },
+      "outputs": [],
+      "source": [
+        "# Cloud project id.\n",
+        "PROJECT_ID = \"\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# The region you want to launch jobs in.\n",
+        "REGION = \"us-central1\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# The Cloud Storage bucket for storing experiments output. Fill it without the 'gs://' prefix.\n",
+        "GCS_BUCKET = \"\"  # @param {type:\"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "92f16e22c20b"
+      },
+      "source": [
+        "Initialize Vertex AI API:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "1680c257acfb"
+      },
+      "outputs": [],
+      "source": [
+        "from google.cloud import aiplatform\n",
+        "\n",
+        "aiplatform.init(project=PROJECT_ID, location=REGION, staging_bucket=GCS_BUCKET)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "6ca48b699d17"
+      },
+      "source": [
+        "### Define constants"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "de9882ea89ea"
+      },
+      "outputs": [],
+      "source": [
+        "# The pre-built serving docker image. It contains serving scripts and models.\n",
+        "SERVE_DOCKER_URI = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/pytorch-transformers-serve:latest\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "10188266a5cd"
+      },
+      "source": [
+        "### Define common functions"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "cac4478ae098"
+      },
+      "outputs": [],
+      "source": [
+        "import base64\n",
+        "import os\n",
+        "from datetime import datetime\n",
+        "from io import BytesIO\n",
+        "\n",
+        "import requests\n",
+        "from google.cloud import aiplatform\n",
+        "from PIL import Image\n",
+        "\n",
+        "\n",
+        "def create_job_name(prefix):\n",
+        "    user = os.environ.get(\"USER\")\n",
+        "    now = datetime.now().strftime(\"%Y%m%d_%H%M%S\")\n",
+        "    job_name = f\"{prefix}-{user}-{now}\"\n",
+        "    return job_name\n",
+        "\n",
+        "\n",
+        "def download_image(url):\n",
+        "    response = requests.get(url)\n",
+        "    return Image.open(BytesIO(response.content))\n",
+        "\n",
+        "\n",
+        "def image_to_base64(image, format=\"JPEG\"):\n",
+        "    buffer = BytesIO()\n",
+        "    image.save(buffer, format=format)\n",
+        "    image_str = base64.b64encode(buffer.getvalue()).decode(\"utf-8\")\n",
+        "    return image_str\n",
+        "\n",
+        "\n",
+        "def base64_to_image(image_str):\n",
+        "    image = Image.open(BytesIO(base64.b64decode(image_str)))\n",
+        "    return image\n",
+        "\n",
+        "\n",
+        "def image_grid(imgs, rows=2, cols=2):\n",
+        "    w, h = imgs[0].size\n",
+        "    grid = Image.new(\"RGB\", size=(cols * w, rows * h))\n",
+        "    for i, img in enumerate(imgs):\n",
+        "        grid.paste(img, box=(i % cols * w, i // cols * h))\n",
+        "    return grid\n",
+        "\n",
+        "\n",
+        "def deploy_model(model_id, task):\n",
+        "    model_name = \"clip\"\n",
+        "    endpoint = aiplatform.Endpoint.create(display_name=f\"{model_name}-endpoint\")\n",
+        "    serving_env = {\n",
+        "        \"MODEL_ID\": model_id,\n",
+        "        \"TASK\": task,\n",
+        "    }\n",
+        "    # If the model_id is a GCS path, use artifact_uri to pass it to serving docker.\n",
+        "    artifact_uri = model_id if model_id.startswith(\"gs://\") else None\n",
+        "    model = aiplatform.Model.upload(\n",
+        "        display_name=model_name,\n",
+        "        serving_container_image_uri=SERVE_DOCKER_URI,\n",
+        "        serving_container_ports=[7080],\n",
+        "        serving_container_predict_route=\"/predictions/diffusers_serving\",\n",
+        "        serving_container_health_route=\"/ping\",\n",
+        "        serving_container_environment_variables=serving_env,\n",
+        "        artifact_uri=artifact_uri,\n",
+        "    )\n",
+        "    model.deploy(\n",
+        "        endpoint=endpoint,\n",
+        "        machine_type=\"n1-standard-8\",\n",
+        "        accelerator_type=\"NVIDIA_TESLA_T4\",\n",
+        "        accelerator_count=1,\n",
+        "        deploy_request_timeout=1800,\n",
+        "    )\n",
+        "    return model, endpoint"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d2d72ecdb8c9"
+      },
+      "source": [
+        "## Upload and deploy models"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "9448c5f545fa"
+      },
+      "source": [
+        "This section uploads the pre-trained model to Model Registry and deploys it on the Endpoint with 1 T4 GPU.\n",
+        "\n",
+        "The model deployment step will take ~15 minutes to complete.\n",
+        "\n",
+        "Once deployed, you can send images and object texts to get classification results."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b4b46c28d8b1"
+      },
+      "outputs": [],
+      "source": [
+        "model, endpoint = deploy_model(\n",
+        "    model_id=\"openai/clip-vit-base-patch32\", task=\"zero-shot-image-classification\"\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "6be655247cb1"
+      },
+      "outputs": [],
+      "source": [
+        "image1 = download_image(\"http://images.cocodataset.org/val2017/000000039769.jpg\")\n",
+        "image2 = download_image(\"http://images.cocodataset.org/val2017/000000000285.jpg\")\n",
+        "grid = image_grid([image1, image2], 1, 2)\n",
+        "display(grid)\n",
+        "\n",
+        "instances = [\n",
+        "    {\"image\": image_to_base64(image1), \"text\": \"two cats\"},\n",
+        "    {\"image\": image_to_base64(image2), \"text\": \"a bear\"},\n",
+        "]\n",
+        "preds = endpoint.predict(instances=instances).predictions\n",
+        "print(preds)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "db7ffebdb4be"
+      },
+      "source": [
+        "### Clean up resources"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "2ccf3714dbe9"
+      },
+      "outputs": [],
+      "source": [
+        "# Delete models.\n",
+        "model.delete()\n",
+        "\n",
+        "# Undeploy model and delete endpoint.\n",
+        "endpoint.delete(force=True)"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "name": "model_garden_pytorch_clip.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/community/model_garden/model_garden_pytorch_controlnet.ipynb

@@ -0,0 +1,622 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "7d9bbf86da5e"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2023 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "99c1c3fc2ca5"
+      },
+      "source": [
+        "# Vertex AI Model Garden - ControlNet\n",
+        "\n",
+        "<table align=\"left\">\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_controlnet.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_controlnet.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>                                                                                               <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/notebooks/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_controlnet.ipynb\">\n",
+        "      <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "3de7470326a2"
+      },
+      "source": [
+        "## Overview\n",
+        "\n",
+        "This notebook demonstrates finetuning the [ControlNet](https://huggingface.co/lllyasviel/ControlNet) with the [fusing/fill50k](https://huggingface.co/datasets/fusing/fill50k) dataset and deploying the model on Vertex AI for online prediction.\n",
+        "\n",
+        "### Objective\n",
+        "\n",
+        "- Finetune the ControlNet model.\n",
+        "- Upload the model to [Model Registry](https://cloud.google.com/vertex-ai/docs/model-registry/introduction).\n",
+        "- Deploy the model on [Endpoint](https://cloud.google.com/vertex-ai/docs/predictions/using-private-endpoints).\n",
+        "- Run online predictions for text-guided-image-to-image.\n",
+        "\n",
+        "### Costs\n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI\n",
+        "* Cloud Storage\n",
+        "\n",
+        "Learn about [Vertex AI pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage pricing](https://cloud.google.com/storage/pricing), and use the [Pricing Calculator](https://cloud.google.com/products/calculator/) to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "264c07757582"
+      },
+      "source": [
+        "## Setup environment\n",
+        "\n",
+        "**NOTE**: Jupyter runs lines prefixed with `!` as shell commands, and it interpolates Python variables prefixed with `$` into these commands."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d73ffa0c0b83"
+      },
+      "source": [
+        "### Colab only"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "2707b02ef5df"
+      },
+      "outputs": [],
+      "source": [
+        "!pip3 install --upgrade google-cloud-aiplatform"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b60a4d7100bf"
+      },
+      "outputs": [],
+      "source": [
+        "from google.colab import auth as google_auth\n",
+        "\n",
+        "google_auth.authenticate_user()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "fb671e75ca7b"
+      },
+      "source": [
+        "### Install dependencies"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "dc8ee367fb42"
+      },
+      "outputs": [],
+      "source": [
+        "# Install gdown for downloading example training images.\n",
+        "!pip install gdown\n",
+        "# Install libs for generating conditioning images for ControlNet.\n",
+        "!pip install opencv-python"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "5244aac3d929"
+      },
+      "source": [
+        "Restart the notebook kernel after installs."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "567212ff53a6"
+      },
+      "outputs": [],
+      "source": [
+        "import IPython\n",
+        "\n",
+        "app = IPython.Application.instance()\n",
+        "app.kernel.do_shutdown(True)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "bb7adab99e41"
+      },
+      "source": [
+        "### Setup Google Cloud project\n",
+        "\n",
+        "1. [Select or create a Google Cloud project](https://console.cloud.google.com/cloud-resource-manager). When you first create an account, you get a $300 free credit towards your compute/storage costs.\n",
+        "\n",
+        "1. [Make sure that billing is enabled for your project](https://cloud.google.com/billing/docs/how-to/modify-project).\n",
+        "\n",
+        "1. [Enable the Vertex AI API and Compute Engine API](https://console.cloud.google.com/flows/enableapi?apiid=aiplatform.googleapis.com,compute_component).\n",
+        "\n",
+        "1. [Create a Cloud Storage bucket](https://cloud.google.com/storage/docs/creating-buckets) for storing experiment outputs."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "6c460088b873"
+      },
+      "source": [
+        "Fill following variables for experiments environment:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "855d6b96f291"
+      },
+      "outputs": [],
+      "source": [
+        "# Cloud project id.\n",
+        "PROJECT_ID = \"\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# The region you want to launch jobs in.\n",
+        "REGION = \"us-central1\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# The Cloud Storage bucket for storing experiments output. Fill it without the 'gs://' prefix.\n",
+        "GCS_BUCKET = \"\"  # @param {type:\"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "e828eb320337"
+      },
+      "source": [
+        "Initialize Vertex AI API:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "12cd25839741"
+      },
+      "outputs": [],
+      "source": [
+        "from google.cloud import aiplatform\n",
+        "\n",
+        "aiplatform.init(project=PROJECT_ID, location=REGION, staging_bucket=GCS_BUCKET)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "2cc825514deb"
+      },
+      "source": [
+        "### Define constants"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b42bd4fa2b2d"
+      },
+      "outputs": [],
+      "source": [
+        "# The pre-built training docker image. It contains training scripts and models.\n",
+        "TRAIN_DOCKER_URI = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/pytorch-diffusers-train:latest\"\n",
+        "\n",
+        "# The pre-built serving docker image. It contains serving scripts and models.\n",
+        "SERVE_DOCKER_URI = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/pytorch-diffusers-serve:latest\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0c250872074f"
+      },
+      "source": [
+        "### Define common functions"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 1,
+      "metadata": {
+        "id": "354da31189dc"
+      },
+      "outputs": [],
+      "source": [
+        "import base64\n",
+        "from io import BytesIO\n",
+        "\n",
+        "import cv2\n",
+        "import numpy as np\n",
+        "import requests\n",
+        "from PIL import Image\n",
+        "\n",
+        "\n",
+        "def create_job_name(prefix):\n",
+        "    user = os.environ.get(\"USER\")\n",
+        "    now = datetime.now().strftime(\"%Y%m%d_%H%M%S\")\n",
+        "    job_name = f\"{prefix}-{user}-{now}\"\n",
+        "    return job_name\n",
+        "\n",
+        "\n",
+        "def download_image(url):\n",
+        "    response = requests.get(url)\n",
+        "    return Image.open(BytesIO(response.content))\n",
+        "\n",
+        "\n",
+        "def image_to_base64(image, format=\"JPEG\"):\n",
+        "    buffer = BytesIO()\n",
+        "    image.save(buffer, format=format)\n",
+        "    image_str = base64.b64encode(buffer.getvalue()).decode(\"utf-8\")\n",
+        "    return image_str\n",
+        "\n",
+        "\n",
+        "def base64_to_image(image_str):\n",
+        "    image = Image.open(BytesIO(base64.b64decode(image_str)))\n",
+        "    return image\n",
+        "\n",
+        "\n",
+        "def image_grid(imgs, rows=2, cols=2):\n",
+        "    w, h = imgs[0].size\n",
+        "    grid = Image.new(\"RGB\", size=(cols * w, rows * h))\n",
+        "    for i, img in enumerate(imgs):\n",
+        "        grid.paste(img, box=(i % cols * w, i // cols * h))\n",
+        "    return grid\n",
+        "\n",
+        "\n",
+        "def canny(image):\n",
+        "    image = np.array(image)\n",
+        "    image = cv2.Canny(image, 100, 200)\n",
+        "    image = image[:, :, None]\n",
+        "    image = np.concatenate([image, image, image], axis=2)\n",
+        "    image = Image.fromarray(image)\n",
+        "    return image\n",
+        "\n",
+        "\n",
+        "def deploy_model(model_id, task):\n",
+        "    model_name = \"controlnet\"\n",
+        "    endpoint = aiplatform.Endpoint.create(display_name=f\"{model_name}-endpoint\")\n",
+        "    serving_env = {\n",
+        "        \"MODEL_ID\": model_id,\n",
+        "        \"TASK\": task,\n",
+        "    }\n",
+        "    # If the model_id is a GCS path, use artifact_uri to pass it to serving docker.\n",
+        "    artifact_uri = model_id if model_id.startswith(\"gs://\") else None\n",
+        "    model = aiplatform.Model.upload(\n",
+        "        display_name=model_name,\n",
+        "        serving_container_image_uri=SERVE_DOCKER_URI,\n",
+        "        serving_container_ports=[7080],\n",
+        "        serving_container_predict_route=\"/predictions/diffusers_serving\",\n",
+        "        serving_container_health_route=\"/ping\",\n",
+        "        serving_container_environment_variables=serving_env,\n",
+        "        artifact_uri=artifact_uri,\n",
+        "    )\n",
+        "    model.deploy(\n",
+        "        endpoint=endpoint,\n",
+        "        machine_type=\"n1-standard-8\",\n",
+        "        accelerator_type=\"NVIDIA_TESLA_V100\",\n",
+        "        accelerator_count=1,\n",
+        "        deploy_request_timeout=1800,\n",
+        "    )\n",
+        "    return model, endpoint"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "e70e3519ff8b"
+      },
+      "source": [
+        "## Finetune with fill50k dataset"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0dc65d8f0689"
+      },
+      "source": [
+        "This section uses the [fusing/fill50k](https://huggingface.co/datasets/fusing/fill50k) dataset to finetune the ControlNet model.\n",
+        "\n",
+        "The job will run on 1 A100 GPU and take ~7 hours to finish 1 epoch of training.\n",
+        "\n",
+        "The ControlNet model will be saved after the finetuning job finishs and it can be loaded to run inference later."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "65467b361315"
+      },
+      "outputs": [],
+      "source": [
+        "# The pre-trained stable diffusion model to be loaded.\n",
+        "stable_diffusion_model_id = \"runwayml/stable-diffusion-v1-5\"\n",
+        "# The datase id to be loaded.\n",
+        "dataset_id = \"fusing/fill50k\"\n",
+        "# The output path.\n",
+        "output_dir = f\"/gcs/{GCS_BUCKET}/controlnet/output\"\n",
+        "\n",
+        "# Worker pool spec.\n",
+        "machine_type = \"a2-highgpu-1g\"\n",
+        "num_nodes = 1\n",
+        "gpu_type = \"NVIDIA_TESLA_A100\"\n",
+        "num_gpus = 1\n",
+        "\n",
+        "# Setup training job.\n",
+        "job_name = create_job_name(\"controlnet\")\n",
+        "job = aiplatform.CustomContainerTrainingJob(\n",
+        "    display_name=job_name,\n",
+        "    container_uri=TRAIN_DOCKER_URI,\n",
+        ")\n",
+        "\n",
+        "# Pass training arguments and launch job.\n",
+        "# See https://github.com/huggingface/diffusers/blob/main/examples/controlnet/train_controlnet.py\n",
+        "# for a full list of training arguments.\n",
+        "model = job.run(\n",
+        "    args=[\n",
+        "        \"controlnet/train_controlnet.py\",\n",
+        "        \"--tracker_project_name=train_controlnet\",\n",
+        "        f\"--pretrained_model_name_or_path={stable_diffusion_model_id}\",\n",
+        "        f\"--output_dir={output_dir}\",\n",
+        "        f\"--dataset_name={dataset_id}\",\n",
+        "        \"--resolution=512\",\n",
+        "        \"--learning_rate=1e-5\",\n",
+        "        \"--train_batch_size=2\",\n",
+        "    ],\n",
+        "    replica_count=num_nodes,\n",
+        "    machine_type=machine_type,\n",
+        "    accelerator_type=gpu_type,\n",
+        "    accelerator_count=num_gpus,\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "bf7f82732e61"
+      },
+      "source": [
+        "## Upload and Deploy models"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "1cc26e68d7b0"
+      },
+      "source": [
+        "This section uploads the model to Model Registry and deploys it on the Endpoint.\n",
+        "\n",
+        "The model deployment step will take ~15 minutes to complete."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "cd7b56421392"
+      },
+      "source": [
+        "### Pre-trained canny model"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "6d331b1ea337"
+      },
+      "source": [
+        "Deploy the pre-trained [lllyasviel/sd-controlnet-canny](https://huggingface.co/lllyasviel/sd-controlnet-canny) model for the text-guided image-to-image task. When deployed on one V100 GPU, the average inference time of a request is ~15 seconds."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "bf55e38815dc"
+      },
+      "outputs": [],
+      "source": [
+        "model, endpoint = deploy_model(\n",
+        "    model_id=\"lllyasviel/sd-controlnet-canny\", task=\"controlnet\"\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "4ab04da3ec9a"
+      },
+      "outputs": [],
+      "source": [
+        "init_image = download_image(\n",
+        "    \"https://huggingface.co/takuma104/controlnet_dev/resolve/main/gen_compare/output_images/diffusers/output_bird_canny_1.png\"\n",
+        ")\n",
+        "display(init_image)\n",
+        "image = canny(init_image)\n",
+        "display(image)\n",
+        "\n",
+        "instances = [\n",
+        "    {\n",
+        "        \"prompt\": \"bird\",\n",
+        "        \"image\": image_to_base64(image),\n",
+        "    },\n",
+        "]\n",
+        "response = endpoint.predict(instances=instances)\n",
+        "images = [base64_to_image(image) for image in response.predictions]\n",
+        "display(images[0])"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "af21a3cff1e0"
+      },
+      "source": [
+        "Clean up resources:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "911406c1561e"
+      },
+      "outputs": [],
+      "source": [
+        "# Delete models.\n",
+        "model.delete()\n",
+        "\n",
+        "# Undeploy model and delete endpoint.\n",
+        "endpoint.delete(force=True)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "c1e51f764a60"
+      },
+      "source": [
+        "### Custom finetuned fill50k model"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "fa686a54047c"
+      },
+      "source": [
+        "Deploy the finetuned fill50k model above for the text-guided image-to-image task. When deployed on one V100 GPU, the averaged inference time of a request is ~15 seconds."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "65e32356fbd1"
+      },
+      "outputs": [],
+      "source": [
+        "model, endpoint = deploy_model(\n",
+        "    model_id=f\"gs://{GCS_BUCKET}/controlnet/output\", task=\"image-to-image\"\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "83a50fd4a1ed"
+      },
+      "outputs": [],
+      "source": [
+        "init_image = download_image(\n",
+        "    \"https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/controlnet_training/conditioning_image_1.png\"\n",
+        ")\n",
+        "display(init_image)\n",
+        "\n",
+        "instances = [\n",
+        "    {\n",
+        "        \"prompt\": \"red circle with green background\",\n",
+        "        \"image\": image_to_base64(init_image, format=\"PNG\"),\n",
+        "    },\n",
+        "]\n",
+        "response = endpoint.predict(instances=instances)\n",
+        "images = [base64_to_image(image) for image in response.predictions]\n",
+        "display(images[0])"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "ed3795d474b9"
+      },
+      "source": [
+        "Clean up resources:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b53b883257b4"
+      },
+      "outputs": [],
+      "source": [
+        "# Delete models.\n",
+        "model.delete()\n",
+        "\n",
+        "# Undeploy model and delete endpoint.\n",
+        "endpoint.delete(force=True)"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "name": "model_garden_pytorch_controlnet.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/community/model_garden/model_garden_pytorch_instructpix2pix.ipynb

@@ -0,0 +1,415 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "7d9bbf86da5e"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2023 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "2bd716bf3e39"
+      },
+      "source": [
+        "# Vertex AI Model Garden - InstructPix2Pix\n",
+        "\n",
+        "<table align=\"left\">\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_instructpix2pix.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_instructpix2pix.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>                                                                                               <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/notebooks/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_instructpix2pix.ipynb\">\n",
+        "      <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d8cd12648da4"
+      },
+      "source": [
+        "## Overview\n",
+        "\n",
+        "This notebook demonstrates deploying the pre-trained [InstructPix2Pix](https://huggingface.co/timbrooks/instruct-pix2pix) model on Vertex AI for online prediction.\n",
+        "\n",
+        "### Objective\n",
+        "\n",
+        "- Upload the model to [Model Registry](https://cloud.google.com/vertex-ai/docs/model-registry/introduction).\n",
+        "- Deploy the model on [Endpoint](https://cloud.google.com/vertex-ai/docs/predictions/using-private-endpoints).\n",
+        "- Run online predictions for text-guided image-to-image.\n",
+        "\n",
+        "### Costs\n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI\n",
+        "* Cloud Storage\n",
+        "\n",
+        "Learn about [Vertex AI pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage pricing](https://cloud.google.com/storage/pricing), and use the [Pricing Calculator](https://cloud.google.com/products/calculator/) to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "264c07757582"
+      },
+      "source": [
+        "## Setup environment\n",
+        "\n",
+        "**NOTE**: Jupyter runs lines prefixed with `!` as shell commands, and it interpolates Python variables prefixed with `$` into these commands."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d73ffa0c0b83"
+      },
+      "source": [
+        "### Colab only"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "2707b02ef5df"
+      },
+      "outputs": [],
+      "source": [
+        "!pip3 install --upgrade google-cloud-aiplatform"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b60a4d7100bf"
+      },
+      "outputs": [],
+      "source": [
+        "from google.colab import auth as google_auth\n",
+        "\n",
+        "google_auth.authenticate_user()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0f826ff482a2"
+      },
+      "source": [
+        "### Setup Google Cloud project\n",
+        "\n",
+        "1. [Select or create a Google Cloud project](https://console.cloud.google.com/cloud-resource-manager). When you first create an account, you get a $300 free credit towards your compute/storage costs.\n",
+        "\n",
+        "1. [Make sure that billing is enabled for your project](https://cloud.google.com/billing/docs/how-to/modify-project).\n",
+        "\n",
+        "1. [Enable the Vertex AI API and Compute Engine API](https://console.cloud.google.com/flows/enableapi?apiid=aiplatform.googleapis.com,compute_component).\n",
+        "\n",
+        "1. [Create a Cloud Storage bucket](https://cloud.google.com/storage/docs/creating-buckets) for storing experiment outputs."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "8958ebc71868"
+      },
+      "source": [
+        "Fill following variables for experiments environment:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "9db30f827a65"
+      },
+      "outputs": [],
+      "source": [
+        "# Cloud project id.\n",
+        "PROJECT_ID = \"\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# The region you want to launch jobs in.\n",
+        "REGION = \"us-central1\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# The Cloud Storage bucket for storing experiments output. Fill it without the 'gs://' prefix.\n",
+        "GCS_BUCKET = \"\"  # @param {type:\"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "92f16e22c20b"
+      },
+      "source": [
+        "Initialize Vertex AI API:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "1680c257acfb"
+      },
+      "outputs": [],
+      "source": [
+        "from google.cloud import aiplatform\n",
+        "\n",
+        "aiplatform.init(project=PROJECT_ID, location=REGION, staging_bucket=GCS_BUCKET)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "6ca48b699d17"
+      },
+      "source": [
+        "### Define constants"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "de9882ea89ea"
+      },
+      "outputs": [],
+      "source": [
+        "# The pre-built serving docker image. It contains serving scripts and models.\n",
+        "SERVE_DOCKER_URI = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/pytorch-diffusers-serve:latest\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "10188266a5cd"
+      },
+      "source": [
+        "### Define common functions"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "cac4478ae098"
+      },
+      "outputs": [],
+      "source": [
+        "import base64\n",
+        "import os\n",
+        "from datetime import datetime\n",
+        "from io import BytesIO\n",
+        "\n",
+        "import requests\n",
+        "from google.cloud import aiplatform\n",
+        "from PIL import Image\n",
+        "\n",
+        "\n",
+        "def create_job_name(prefix):\n",
+        "    user = os.environ.get(\"USER\")\n",
+        "    now = datetime.now().strftime(\"%Y%m%d_%H%M%S\")\n",
+        "    job_name = f\"{prefix}-{user}-{now}\"\n",
+        "    return job_name\n",
+        "\n",
+        "\n",
+        "def download_image(url):\n",
+        "    response = requests.get(url)\n",
+        "    return Image.open(BytesIO(response.content))\n",
+        "\n",
+        "\n",
+        "def image_to_base64(image, format=\"JPEG\"):\n",
+        "    buffer = BytesIO()\n",
+        "    image.save(buffer, format=format)\n",
+        "    image_str = base64.b64encode(buffer.getvalue()).decode(\"utf-8\")\n",
+        "    return image_str\n",
+        "\n",
+        "\n",
+        "def base64_to_image(image_str):\n",
+        "    image = Image.open(BytesIO(base64.b64decode(image_str)))\n",
+        "    return image\n",
+        "\n",
+        "\n",
+        "def image_grid(imgs, rows=2, cols=2):\n",
+        "    w, h = imgs[0].size\n",
+        "    grid = Image.new(\"RGB\", size=(cols * w, rows * h))\n",
+        "    for i, img in enumerate(imgs):\n",
+        "        grid.paste(img, box=(i % cols * w, i // cols * h))\n",
+        "    return grid\n",
+        "\n",
+        "\n",
+        "def deploy_model(model_id, task):\n",
+        "    model_name = \"instruct-pix2pix\"\n",
+        "    endpoint = aiplatform.Endpoint.create(display_name=f\"{model_name}-endpoint\")\n",
+        "    serving_env = {\n",
+        "        \"MODEL_ID\": model_id,\n",
+        "        \"TASK\": task,\n",
+        "    }\n",
+        "    # If the model_id is a GCS path, use artifact_uri to pass it to serving docker.\n",
+        "    artifact_uri = model_id if model_id.startswith(\"gs://\") else None\n",
+        "    model = aiplatform.Model.upload(\n",
+        "        display_name=model_name,\n",
+        "        serving_container_image_uri=SERVE_DOCKER_URI,\n",
+        "        serving_container_ports=[7080],\n",
+        "        serving_container_predict_route=\"/predictions/diffusers_serving\",\n",
+        "        serving_container_health_route=\"/ping\",\n",
+        "        serving_container_environment_variables=serving_env,\n",
+        "        artifact_uri=artifact_uri,\n",
+        "    )\n",
+        "    model.deploy(\n",
+        "        endpoint=endpoint,\n",
+        "        machine_type=\"n1-standard-8\",\n",
+        "        accelerator_type=\"NVIDIA_TESLA_V100\",\n",
+        "        accelerator_count=1,\n",
+        "        deploy_request_timeout=1800,\n",
+        "    )\n",
+        "    return model, endpoint"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d2d72ecdb8c9"
+      },
+      "source": [
+        "## Upload and deploy models"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "9448c5f545fa"
+      },
+      "source": [
+        "This section uploads the pre-trained model to Model Registry and deploys it on the Endpoint.\n",
+        "\n",
+        "The model deployment step will take ~15 minutes to complete."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "c277da31bde6"
+      },
+      "source": [
+        "### Text-guided image-to-image"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "a5a86996222c"
+      },
+      "source": [
+        "Deploy the InstructPix2Pix model for the text-guided image-to-image task.\n",
+        "\n",
+        "Once deployed, you can send prompts to the endpoint to generated images.\n",
+        "\n",
+        "When deployed on one V100 GPU, the averaged inference time of a request is ~15 seconds."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b4b46c28d8b1"
+      },
+      "outputs": [],
+      "source": [
+        "model, endpoint = deploy_model(\n",
+        "    model_id=\"timbrooks/instruct-pix2pix\", task=\"instruct-pix2pix\"\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "6be655247cb1"
+      },
+      "outputs": [],
+      "source": [
+        "init_image = download_image(\n",
+        "    \"https://huggingface.co/datasets/diffusers/diffusers-images-docs/resolve/main/mountain.png\"\n",
+        ")\n",
+        "display(init_image)\n",
+        "instances = [\n",
+        "    {\n",
+        "        \"prompt\": \"Add fire to the mountain\",\n",
+        "        \"image\": image_to_base64(init_image),\n",
+        "    },\n",
+        "]\n",
+        "response = endpoint.predict(instances=instances)\n",
+        "images = [base64_to_image(image) for image in response.predictions]\n",
+        "display(images[0])"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "db7ffebdb4be"
+      },
+      "source": [
+        "### Clean up resources"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "2ccf3714dbe9"
+      },
+      "outputs": [],
+      "source": [
+        "# Delete models.\n",
+        "model.delete()\n",
+        "\n",
+        "# Undeploy model and delete endpoint.\n",
+        "endpoint.delete(force=True)"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "name": "model_garden_pytorch_instructpix2pix.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/community/model_garden/model_garden_pytorch_layoutml_document_qa.ipynb

@@ -0,0 +1,394 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "7d9bbf86da5e"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2023 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "2bd716bf3e39"
+      },
+      "source": [
+        "# Vertex AI Model Garden - LayoutML Document QA\n",
+        "\n",
+        "<table align=\"left\">\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_layoutml_document_qa.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_layoutml_document_qa.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>                                                                                               <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/notebooks/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_layoutml_document_qa.ipynb\">\n",
+        "      <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d8cd12648da4"
+      },
+      "source": [
+        "## Overview\n",
+        "\n",
+        "This notebook demonstrates deploying the pre-trained [LayoutML](https://huggingface.co/impira/layoutlm-document-qa) model on Vertex AI for online prediction.\n",
+        "\n",
+        "### Objective\n",
+        "\n",
+        "- Upload the model to [Model Registry](https://cloud.google.com/vertex-ai/docs/model-registry/introduction).\n",
+        "- Deploy the model on [Endpoint](https://cloud.google.com/vertex-ai/docs/predictions/using-private-endpoints).\n",
+        "- Run online predictions for image captioning.\n",
+        "\n",
+        "### Costs\n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI\n",
+        "* Cloud Storage\n",
+        "\n",
+        "Learn about [Vertex AI pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage pricing](https://cloud.google.com/storage/pricing), and use the [Pricing Calculator](https://cloud.google.com/products/calculator/) to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "264c07757582"
+      },
+      "source": [
+        "## Setup environment\n",
+        "\n",
+        "**NOTE**: Jupyter runs lines prefixed with `!` as shell commands, and it interpolates Python variables prefixed with `$` into these commands."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d73ffa0c0b83"
+      },
+      "source": [
+        "### Colab only"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "2707b02ef5df"
+      },
+      "outputs": [],
+      "source": [
+        "!pip3 install --upgrade google-cloud-aiplatform"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b60a4d7100bf"
+      },
+      "outputs": [],
+      "source": [
+        "from google.colab import auth as google_auth\n",
+        "\n",
+        "google_auth.authenticate_user()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0f826ff482a2"
+      },
+      "source": [
+        "### Setup Google Cloud project\n",
+        "\n",
+        "1. [Select or create a Google Cloud project](https://console.cloud.google.com/cloud-resource-manager). When you first create an account, you get a $300 free credit towards your compute/storage costs.\n",
+        "\n",
+        "1. [Make sure that billing is enabled for your project](https://cloud.google.com/billing/docs/how-to/modify-project).\n",
+        "\n",
+        "1. [Enable the Vertex AI API and Compute Engine API](https://console.cloud.google.com/flows/enableapi?apiid=aiplatform.googleapis.com,compute_component).\n",
+        "\n",
+        "1. [Create a Cloud Storage bucket](https://cloud.google.com/storage/docs/creating-buckets) for storing experiment outputs."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "8958ebc71868"
+      },
+      "source": [
+        "Fill following variables for experiments environment:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "9db30f827a65"
+      },
+      "outputs": [],
+      "source": [
+        "# Cloud project id.\n",
+        "PROJECT_ID = \"\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# The region you want to launch jobs in.\n",
+        "REGION = \"us-central1\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# The Cloud Storage bucket for storing experiments output. Fill it without the 'gs://' prefix.\n",
+        "GCS_BUCKET = \"\"  # @param {type:\"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "92f16e22c20b"
+      },
+      "source": [
+        "Initialize Vertex AI API:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "1680c257acfb"
+      },
+      "outputs": [],
+      "source": [
+        "from google.cloud import aiplatform\n",
+        "\n",
+        "aiplatform.init(project=PROJECT_ID, location=REGION, staging_bucket=GCS_BUCKET)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "6ca48b699d17"
+      },
+      "source": [
+        "### Define constants"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "de9882ea89ea"
+      },
+      "outputs": [],
+      "source": [
+        "# The pre-built serving docker image. It contains serving scripts and models.\n",
+        "SERVE_DOCKER_URI = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/pytorch-transformers-serve:latest\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "10188266a5cd"
+      },
+      "source": [
+        "### Define common functions"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "cac4478ae098"
+      },
+      "outputs": [],
+      "source": [
+        "import base64\n",
+        "import os\n",
+        "from datetime import datetime\n",
+        "from io import BytesIO\n",
+        "\n",
+        "import requests\n",
+        "from google.cloud import aiplatform\n",
+        "from PIL import Image\n",
+        "\n",
+        "\n",
+        "def create_job_name(prefix):\n",
+        "    user = os.environ.get(\"USER\")\n",
+        "    now = datetime.now().strftime(\"%Y%m%d_%H%M%S\")\n",
+        "    job_name = f\"{prefix}-{user}-{now}\"\n",
+        "    return job_name\n",
+        "\n",
+        "\n",
+        "def download_image(url):\n",
+        "    response = requests.get(url)\n",
+        "    return Image.open(BytesIO(response.content))\n",
+        "\n",
+        "\n",
+        "def image_to_base64(image, format=\"JPEG\"):\n",
+        "    buffer = BytesIO()\n",
+        "    image.save(buffer, format=format)\n",
+        "    image_str = base64.b64encode(buffer.getvalue()).decode(\"utf-8\")\n",
+        "    return image_str\n",
+        "\n",
+        "\n",
+        "def base64_to_image(image_str):\n",
+        "    image = Image.open(BytesIO(base64.b64decode(image_str)))\n",
+        "    return image\n",
+        "\n",
+        "\n",
+        "def image_grid(imgs, rows=2, cols=2):\n",
+        "    w, h = imgs[0].size\n",
+        "    grid = Image.new(\"RGB\", size=(cols * w, rows * h))\n",
+        "    for i, img in enumerate(imgs):\n",
+        "        grid.paste(img, box=(i % cols * w, i // cols * h))\n",
+        "    return grid\n",
+        "\n",
+        "\n",
+        "def deploy_model(model_id, task):\n",
+        "    model_name = \"layoutml\"\n",
+        "    endpoint = aiplatform.Endpoint.create(display_name=f\"{model_name}-endpoint\")\n",
+        "    serving_env = {\n",
+        "        \"MODEL_ID\": model_id,\n",
+        "        \"TASK\": task,\n",
+        "    }\n",
+        "    # If the model_id is a GCS path, use artifact_uri to pass it to serving docker.\n",
+        "    artifact_uri = model_id if model_id.startswith(\"gs://\") else None\n",
+        "    model = aiplatform.Model.upload(\n",
+        "        display_name=model_name,\n",
+        "        serving_container_image_uri=SERVE_DOCKER_URI,\n",
+        "        serving_container_ports=[7080],\n",
+        "        serving_container_predict_route=\"/predictions/diffusers_serving\",\n",
+        "        serving_container_health_route=\"/ping\",\n",
+        "        serving_container_environment_variables=serving_env,\n",
+        "        artifact_uri=artifact_uri,\n",
+        "    )\n",
+        "    model.deploy(\n",
+        "        endpoint=endpoint,\n",
+        "        machine_type=\"n1-standard-8\",\n",
+        "        accelerator_type=\"NVIDIA_TESLA_T4\",\n",
+        "        accelerator_count=1,\n",
+        "        deploy_request_timeout=1800,\n",
+        "    )\n",
+        "    return model, endpoint"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d2d72ecdb8c9"
+      },
+      "source": [
+        "## Upload and deploy models"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "9448c5f545fa"
+      },
+      "source": [
+        "This section uploads the pre-trained model to Model Registry and deploys it on the Endpoint with 1 T4 GPU.\n",
+        "\n",
+        "The model deployment step will take ~15 minutes to complete.\n",
+        "\n",
+        "Once deployed, you can send document images and questions to get answers."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b4b46c28d8b1"
+      },
+      "outputs": [],
+      "source": [
+        "model, endpoint = deploy_model(\n",
+        "    model_id=\"impira/layoutlm-document-qa\", task=\"document-question-answering\"\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "6be655247cb1"
+      },
+      "outputs": [],
+      "source": [
+        "image = download_image(\n",
+        "    \"https://huggingface.co/spaces/impira/docquery/resolve/2359223c1837a7587402bda0f2643382a6eefeab/invoice.png\"\n",
+        ")\n",
+        "display(image)\n",
+        "\n",
+        "question = \"What is the name of the signer?\"\n",
+        "instances = [\n",
+        "    {\"image\": image_to_base64(image), \"text\": \"\"},\n",
+        "]\n",
+        "preds = endpoint.predict(instances=instances).predictions\n",
+        "print(question)\n",
+        "print(preds)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "db7ffebdb4be"
+      },
+      "source": [
+        "### Clean up resources"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "2ccf3714dbe9"
+      },
+      "outputs": [],
+      "source": [
+        "# Delete models.\n",
+        "model.delete()\n",
+        "\n",
+        "# Undeploy model and delete endpoint.\n",
+        "endpoint.delete(force=True)"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "name": "model_garden_pytorch_layoutml_document_qa.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/community/model_garden/model_garden_pytorch_owlvit.ipynb

@@ -0,0 +1,390 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "7d9bbf86da5e"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2023 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "2bd716bf3e39"
+      },
+      "source": [
+        "# Vertex AI Model Garden - OWL-ViT\n",
+        "\n",
+        "<table align=\"left\">\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_owlvit.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_owlvit.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>                                                                                               <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/notebooks/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_owlvit.ipynb\">\n",
+        "      <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d8cd12648da4"
+      },
+      "source": [
+        "## Overview\n",
+        "\n",
+        "This notebook demonstrates deploying the pre-trained [OWL-ViT](https://huggingface.co/google/owlvit-base-patch32) model on Vertex AI for online prediction.\n",
+        "\n",
+        "### Objective\n",
+        "\n",
+        "- Upload the model to [Model Registry](https://cloud.google.com/vertex-ai/docs/model-registry/introduction).\n",
+        "- Deploy the model on [Endpoint](https://cloud.google.com/vertex-ai/docs/predictions/using-private-endpoints).\n",
+        "- Run online predictions for image captioning.\n",
+        "\n",
+        "### Costs\n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI\n",
+        "* Cloud Storage\n",
+        "\n",
+        "Learn about [Vertex AI pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage pricing](https://cloud.google.com/storage/pricing), and use the [Pricing Calculator](https://cloud.google.com/products/calculator/) to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "264c07757582"
+      },
+      "source": [
+        "## Setup environment\n",
+        "\n",
+        "**NOTE**: Jupyter runs lines prefixed with `!` as shell commands, and it interpolates Python variables prefixed with `$` into these commands."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d73ffa0c0b83"
+      },
+      "source": [
+        "### Colab only"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "2707b02ef5df"
+      },
+      "outputs": [],
+      "source": [
+        "!pip3 install --upgrade google-cloud-aiplatform"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b60a4d7100bf"
+      },
+      "outputs": [],
+      "source": [
+        "from google.colab import auth as google_auth\n",
+        "\n",
+        "google_auth.authenticate_user()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0f826ff482a2"
+      },
+      "source": [
+        "### Setup Google Cloud project\n",
+        "\n",
+        "1. [Select or create a Google Cloud project](https://console.cloud.google.com/cloud-resource-manager). When you first create an account, you get a $300 free credit towards your compute/storage costs.\n",
+        "\n",
+        "1. [Make sure that billing is enabled for your project](https://cloud.google.com/billing/docs/how-to/modify-project).\n",
+        "\n",
+        "1. [Enable the Vertex AI API and Compute Engine API](https://console.cloud.google.com/flows/enableapi?apiid=aiplatform.googleapis.com,compute_component).\n",
+        "\n",
+        "1. [Create a Cloud Storage bucket](https://cloud.google.com/storage/docs/creating-buckets) for storing experiment outputs."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "8958ebc71868"
+      },
+      "source": [
+        "Fill following variables for experiments environment:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "9db30f827a65"
+      },
+      "outputs": [],
+      "source": [
+        "# Cloud project id.\n",
+        "PROJECT_ID = \"\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# The region you want to launch jobs in.\n",
+        "REGION = \"us-central1\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# The Cloud Storage bucket for storing experiments output. Fill it without the 'gs://' prefix.\n",
+        "GCS_BUCKET = \"\"  # @param {type:\"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "92f16e22c20b"
+      },
+      "source": [
+        "Initialize Vertex AI API:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "1680c257acfb"
+      },
+      "outputs": [],
+      "source": [
+        "from google.cloud import aiplatform\n",
+        "\n",
+        "aiplatform.init(project=PROJECT_ID, location=REGION, staging_bucket=GCS_BUCKET)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "6ca48b699d17"
+      },
+      "source": [
+        "### Define constants"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "de9882ea89ea"
+      },
+      "outputs": [],
+      "source": [
+        "# The pre-built serving docker image. It contains serving scripts and models.\n",
+        "SERVE_DOCKER_URI = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/pytorch-transformers-serve:latest\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "10188266a5cd"
+      },
+      "source": [
+        "### Define common functions"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "cac4478ae098"
+      },
+      "outputs": [],
+      "source": [
+        "import base64\n",
+        "import os\n",
+        "from datetime import datetime\n",
+        "from io import BytesIO\n",
+        "\n",
+        "import requests\n",
+        "from google.cloud import aiplatform\n",
+        "from PIL import Image\n",
+        "\n",
+        "\n",
+        "def create_job_name(prefix):\n",
+        "    user = os.environ.get(\"USER\")\n",
+        "    now = datetime.now().strftime(\"%Y%m%d_%H%M%S\")\n",
+        "    job_name = f\"{prefix}-{user}-{now}\"\n",
+        "    return job_name\n",
+        "\n",
+        "\n",
+        "def download_image(url):\n",
+        "    response = requests.get(url)\n",
+        "    return Image.open(BytesIO(response.content))\n",
+        "\n",
+        "\n",
+        "def image_to_base64(image, format=\"JPEG\"):\n",
+        "    buffer = BytesIO()\n",
+        "    image.save(buffer, format=format)\n",
+        "    image_str = base64.b64encode(buffer.getvalue()).decode(\"utf-8\")\n",
+        "    return image_str\n",
+        "\n",
+        "\n",
+        "def base64_to_image(image_str):\n",
+        "    image = Image.open(BytesIO(base64.b64decode(image_str)))\n",
+        "    return image\n",
+        "\n",
+        "\n",
+        "def image_grid(imgs, rows=2, cols=2):\n",
+        "    w, h = imgs[0].size\n",
+        "    grid = Image.new(\"RGB\", size=(cols * w, rows * h))\n",
+        "    for i, img in enumerate(imgs):\n",
+        "        grid.paste(img, box=(i % cols * w, i // cols * h))\n",
+        "    return grid\n",
+        "\n",
+        "\n",
+        "def deploy_model(model_id, task):\n",
+        "    model_name = \"owl-vit\"\n",
+        "    endpoint = aiplatform.Endpoint.create(display_name=f\"{model_name}-endpoint\")\n",
+        "    serving_env = {\n",
+        "        \"MODEL_ID\": model_id,\n",
+        "        \"TASK\": task,\n",
+        "    }\n",
+        "    # If the model_id is a GCS path, use artifact_uri to pass it to serving docker.\n",
+        "    artifact_uri = model_id if model_id.startswith(\"gs://\") else None\n",
+        "    model = aiplatform.Model.upload(\n",
+        "        display_name=model_name,\n",
+        "        serving_container_image_uri=SERVE_DOCKER_URI,\n",
+        "        serving_container_ports=[7080],\n",
+        "        serving_container_predict_route=\"/predictions/diffusers_serving\",\n",
+        "        serving_container_health_route=\"/ping\",\n",
+        "        serving_container_environment_variables=serving_env,\n",
+        "        artifact_uri=artifact_uri,\n",
+        "    )\n",
+        "    model.deploy(\n",
+        "        endpoint=endpoint,\n",
+        "        machine_type=\"n1-standard-8\",\n",
+        "        accelerator_type=\"NVIDIA_TESLA_T4\",\n",
+        "        accelerator_count=1,\n",
+        "        deploy_request_timeout=1800,\n",
+        "    )\n",
+        "    return model, endpoint"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d2d72ecdb8c9"
+      },
+      "source": [
+        "## Upload and deploy models"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "9448c5f545fa"
+      },
+      "source": [
+        "This section uploads the pre-trained model to Model Registry and deploys it on the Endpoint with 1 T4 GPU.\n",
+        "\n",
+        "The model deployment step will take ~15 minutes to complete.\n",
+        "\n",
+        "Once deployed, you can send images and object texts to get bounding boxes."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b4b46c28d8b1"
+      },
+      "outputs": [],
+      "source": [
+        "model, endpoint = deploy_model(\n",
+        "    model_id=\"google/owlvit-base-patch32\", task=\"zero-shot-object-detection\"\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "6be655247cb1"
+      },
+      "outputs": [],
+      "source": [
+        "image = download_image(\"http://images.cocodataset.org/val2017/000000039769.jpg\")\n",
+        "\n",
+        "instances = [\n",
+        "    {\"image\": image_to_base64(image), \"text\": \"cat\"},\n",
+        "]\n",
+        "preds = endpoint.predict(instances=instances).predictions\n",
+        "draw_image_with_boxes(image, preds[0][\"boxes\"])\n",
+        "print(preds)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "db7ffebdb4be"
+      },
+      "source": [
+        "### Clean up resources"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "2ccf3714dbe9"
+      },
+      "outputs": [],
+      "source": [
+        "# Delete models.\n",
+        "model.delete()\n",
+        "\n",
+        "# Undeploy model and delete endpoint.\n",
+        "endpoint.delete(force=True)"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "name": "model_garden_pytorch_owlvit.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/community/model_garden/model_garden_pytorch_stable_diffusion.ipynb

@@ -0,0 +1,648 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "7d9bbf86da5e"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2023 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "99c1c3fc2ca5"
+      },
+      "source": [
+        "# Vertex AI Model Garden - Stable Diffusion V1.5\n",
+        "\n",
+        "<table align=\"left\">\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_stable_diffusion.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_stable_diffusion.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>                                                                                               <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/notebooks/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_stable_diffusion.ipynb\">\n",
+        "      <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "3de7470326a2"
+      },
+      "source": [
+        "## Overview\n",
+        "\n",
+        "This notebook demonstrates finetuning [runwayml/stable-diffusion-v1-5](https://huggingface.co/runwayml/stable-diffusion-v1-5) with [Dreambooth](https://huggingface.co/docs/diffusers/training/dreambooth) and deploying it on Vertex AI for online prediction.\n",
+        "\n",
+        "### Objective\n",
+        "\n",
+        "- Finetune the stable-diffusion-v1.5 model with [Dreambooth](https://huggingface.co/docs/diffusers/training/dreambooth).\n",
+        "- Upload the model to [Model Registry](https://cloud.google.com/vertex-ai/docs/model-registry/introduction).\n",
+        "- Deploy the model on [Endpoint](https://cloud.google.com/vertex-ai/docs/predictions/using-private-endpoints).\n",
+        "- Run online predictions for text-to-image and text-guided-image-to-image.\n",
+        "\n",
+        "### Costs\n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI\n",
+        "* Cloud Storage\n",
+        "\n",
+        "Learn about [Vertex AI pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage pricing](https://cloud.google.com/storage/pricing), and use the [Pricing Calculator](https://cloud.google.com/products/calculator/) to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "264c07757582"
+      },
+      "source": [
+        "## Setup environment\n",
+        "\n",
+        "**NOTE**: Jupyter runs lines prefixed with `!` as shell commands, and it interpolates Python variables prefixed with `$` into these commands."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d73ffa0c0b83"
+      },
+      "source": [
+        "### Colab only"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "2707b02ef5df"
+      },
+      "outputs": [],
+      "source": [
+        "!pip3 install --upgrade google-cloud-aiplatform"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b60a4d7100bf"
+      },
+      "outputs": [],
+      "source": [
+        "from google.colab import auth as google_auth\n",
+        "\n",
+        "google_auth.authenticate_user()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "fb671e75ca7b"
+      },
+      "source": [
+        "### Install dependencies"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "dc8ee367fb42"
+      },
+      "outputs": [],
+      "source": [
+        "# Install gdown for downloading example training images.\n",
+        "!pip install gdown\n",
+        "# Install gsutil for downloading/uploading data from/to Cloud Storage buckets.\n",
+        "!pip install gsutil"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "5244aac3d929"
+      },
+      "source": [
+        "Restart the notebook kernel after installs."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "567212ff53a6"
+      },
+      "outputs": [],
+      "source": [
+        "import IPython\n",
+        "\n",
+        "app = IPython.Application.instance()\n",
+        "app.kernel.do_shutdown(True)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "bb7adab99e41"
+      },
+      "source": [
+        "### Setup Google Cloud project\n",
+        "\n",
+        "1. [Select or create a Google Cloud project](https://console.cloud.google.com/cloud-resource-manager). When you first create an account, you get a $300 free credit towards your compute/storage costs.\n",
+        "\n",
+        "1. [Make sure that billing is enabled for your project](https://cloud.google.com/billing/docs/how-to/modify-project).\n",
+        "\n",
+        "1. [Enable the Vertex AI API and Compute Engine API](https://console.cloud.google.com/flows/enableapi?apiid=aiplatform.googleapis.com,compute_component).\n",
+        "\n",
+        "1. [Create a Cloud Storage bucket](https://cloud.google.com/storage/docs/creating-buckets) for storing experiment outputs."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "6c460088b873"
+      },
+      "source": [
+        "Fill following variables for experiments environment:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "855d6b96f291"
+      },
+      "outputs": [],
+      "source": [
+        "# Cloud project id.\n",
+        "PROJECT_ID = \"\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# The region you want to launch jobs in.\n",
+        "REGION = \"us-central1\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# The Cloud Storage bucket for storing experiments output. Fill it without the 'gs://' prefix.\n",
+        "GCS_BUCKET = \"\"  # @param {type:\"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "e828eb320337"
+      },
+      "source": [
+        "Initialize Vertex-AI API:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "12cd25839741"
+      },
+      "outputs": [],
+      "source": [
+        "from google.cloud import aiplatform\n",
+        "\n",
+        "aiplatform.init(project=PROJECT_ID, location=REGION, staging_bucket=GCS_BUCKET)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "2cc825514deb"
+      },
+      "source": [
+        "### Define constants"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b42bd4fa2b2d"
+      },
+      "outputs": [],
+      "source": [
+        "# The pre-built training docker image. It contains training scripts and models.\n",
+        "TRAIN_DOCKER_URI = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/pytorch-diffusers-train:latest\"\n",
+        "\n",
+        "# The pre-built serving docker image. It contains serving scripts and models.\n",
+        "SERVE_DOCKER_URI = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/pytorch-diffusers-serve:latest\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0c250872074f"
+      },
+      "source": [
+        "### Define common functions"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "354da31189dc"
+      },
+      "outputs": [],
+      "source": [
+        "import base64\n",
+        "import os\n",
+        "from datetime import datetime\n",
+        "from io import BytesIO\n",
+        "\n",
+        "import requests\n",
+        "from google.cloud import aiplatform\n",
+        "from PIL import Image\n",
+        "\n",
+        "\n",
+        "def create_job_name(prefix):\n",
+        "    user = os.environ.get(\"USER\")\n",
+        "    now = datetime.now().strftime(\"%Y%m%d_%H%M%S\")\n",
+        "    job_name = f\"{prefix}-{user}-{now}\"\n",
+        "    return job_name\n",
+        "\n",
+        "\n",
+        "def download_image(url):\n",
+        "    response = requests.get(url)\n",
+        "    return Image.open(BytesIO(response.content))\n",
+        "\n",
+        "\n",
+        "def image_to_base64(image, format=\"JPEG\"):\n",
+        "    buffer = BytesIO()\n",
+        "    image.save(buffer, format=format)\n",
+        "    image_str = base64.b64encode(buffer.getvalue()).decode(\"utf-8\")\n",
+        "    return image_str\n",
+        "\n",
+        "\n",
+        "def base64_to_image(image_str):\n",
+        "    image = Image.open(BytesIO(base64.b64decode(image_str)))\n",
+        "    return image\n",
+        "\n",
+        "\n",
+        "def image_grid(imgs, rows=2, cols=2):\n",
+        "    w, h = imgs[0].size\n",
+        "    grid = Image.new(\"RGB\", size=(cols * w, rows * h))\n",
+        "    for i, img in enumerate(imgs):\n",
+        "        grid.paste(img, box=(i % cols * w, i // cols * h))\n",
+        "    return grid\n",
+        "\n",
+        "\n",
+        "def deploy_model(model_id, task):\n",
+        "    model_name = \"stable-diffusion-v1\"\n",
+        "    endpoint = aiplatform.Endpoint.create(display_name=f\"{model_name}-{task}-endpoint\")\n",
+        "    serving_env = {\n",
+        "        \"MODEL_ID\": model_id,\n",
+        "        \"TASK\": task,\n",
+        "    }\n",
+        "    # If the model_id is a GCS path, use artifact_uri to pass it to serving docker.\n",
+        "    artifact_uri = model_id if model_id.startswith(\"gs://\") else None\n",
+        "    model = aiplatform.Model.upload(\n",
+        "        display_name=model_name,\n",
+        "        serving_container_image_uri=SERVE_DOCKER_URI,\n",
+        "        serving_container_ports=[7080],\n",
+        "        serving_container_predict_route=\"/predictions/diffusers_serving\",\n",
+        "        serving_container_health_route=\"/ping\",\n",
+        "        serving_container_environment_variables=serving_env,\n",
+        "        artifact_uri=artifact_uri,\n",
+        "    )\n",
+        "    model.deploy(\n",
+        "        endpoint=endpoint,\n",
+        "        machine_type=\"n1-standard-8\",\n",
+        "        accelerator_type=\"NVIDIA_TESLA_V100\",\n",
+        "        accelerator_count=1,\n",
+        "        deploy_request_timeout=1800,\n",
+        "    )\n",
+        "    return model, endpoint"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "e70e3519ff8b"
+      },
+      "source": [
+        "## Finetune with Dreambooth"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0dc65d8f0689"
+      },
+      "source": [
+        "This section uses [dreambooth](https://dreambooth.github.io/) to finetune the [stable-diffusion-v1.5](https://huggingface.co/runwayml/stable-diffusion-v1-5) model with [5 dog images](https://drive.google.com/drive/folders/1BO_dyz-p65qhBRRMRA4TbZ8qW4rB99JZ) to personalize the text-to-image model.\n",
+        "\n",
+        "It finetunes both text encoder and unet of the stable diffusion model up to 800 steps. The whole finetuning job takes 30 minutes to finish using 1 A100 GPU.\n",
+        "\n",
+        "The full model will be saved after the finetuning job finishs and it can be loaded by the [StableDiffusionPipeline](https://huggingface.co/docs/diffusers/api/pipelines/stable_diffusion/text2img) to run inference."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "34048707df5c"
+      },
+      "outputs": [],
+      "source": [
+        "# Download example training images.\n",
+        "!gdown --folder https://drive.google.com/drive/folders/1BO_dyz-p65qhBRRMRA4TbZ8qW4rB99JZ\n",
+        "\n",
+        "# Upload data to Cloud Storage bucket.\n",
+        "!gsutil -m cp -r dog/* gs://{GCS_BUCKET}/dreambooth/dog/\n",
+        "!gsutil -m cp -r dog/* gs://{GCS_BUCKET}/dreambooth/dog_class/"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "969cfeb79317"
+      },
+      "source": [
+        "**NOTE**: If the upload step fails due to lacking of permission, you need to [grant the Storage Object Admin role](https://cloud.google.com/storage/docs/access-control/using-iam-permissions) for the Cloud account of the notebook."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "65467b361315"
+      },
+      "outputs": [],
+      "source": [
+        "# The pre-trained model to be loaded.\n",
+        "model_id = \"runwayml/stable-diffusion-v1-5\"\n",
+        "\n",
+        "# Input and output path.\n",
+        "instance_dir = f\"/gcs/{GCS_BUCKET}/dreambooth/dog\"\n",
+        "class_dir = f\"/gcs/{GCS_BUCKET}/dreambooth/dog_class\"\n",
+        "output_dir = f\"/gcs/{GCS_BUCKET}/dreambooth/output\"\n",
+        "\n",
+        "# Worker pool spec.\n",
+        "machine_type = \"a2-highgpu-1g\"\n",
+        "num_nodes = 1\n",
+        "gpu_type = \"NVIDIA_TESLA_A100\"\n",
+        "num_gpus = 1\n",
+        "\n",
+        "# Setup training job.\n",
+        "job_name = create_job_name(\"dreambooth-stable-diffusion\")\n",
+        "job = aiplatform.CustomContainerTrainingJob(\n",
+        "    display_name=job_name,\n",
+        "    container_uri=TRAIN_DOCKER_URI,\n",
+        ")\n",
+        "\n",
+        "# Pass training arguments and launch job.\n",
+        "# See https://github.com/huggingface/diffusers/blob/v0.14.0/examples/dreambooth/train_dreambooth.py#L75\n",
+        "# for a full list of training arguments.\n",
+        "model = job.run(\n",
+        "    args=[\n",
+        "        \"dreambooth/train_dreambooth.py\",\n",
+        "        f\"--pretrained_model_name_or_path={model_id}\",\n",
+        "        \"--train_text_encoder\",\n",
+        "        f\"--instance_data_dir={instance_dir}\",\n",
+        "        f\"--class_data_dir={class_dir}\",\n",
+        "        f\"--output_dir={output_dir}\",\n",
+        "        \"--with_prior_preservation\",\n",
+        "        \"--prior_loss_weight=1.0\",\n",
+        "        \"--instance_prompt='a photo of sks dog'\",\n",
+        "        \"--class_prompt='a photo of dog'\",\n",
+        "        \"--resolution=512\",\n",
+        "        \"--train_batch_size=1\",\n",
+        "        \"--gradient_checkpointing\",\n",
+        "        \"--learning_rate=2e-6\",\n",
+        "        \"--lr_scheduler=constant\",\n",
+        "        \"--lr_warmup_steps=0\",\n",
+        "        \"--num_class_images=200\",\n",
+        "        \"--max_train_steps=800\",\n",
+        "    ],\n",
+        "    replica_count=num_nodes,\n",
+        "    machine_type=machine_type,\n",
+        "    accelerator_type=gpu_type,\n",
+        "    accelerator_count=num_gpus,\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "bf7f82732e61"
+      },
+      "source": [
+        "## Upload and Deploy models"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "1cc26e68d7b0"
+      },
+      "source": [
+        "This section uploads the model to Model Registry and deploys it on the Endpoint.\n",
+        "\n",
+        "The model deployment step will take ~15 minutes to complete."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "cd7b56421392"
+      },
+      "source": [
+        "### Text-to-image"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "6d331b1ea337"
+      },
+      "source": [
+        "Deploy the stable diffusion model for the text-to-image task.\n",
+        "\n",
+        "Once deployed, you can send a batch of text prompts to the endpoint to generated images.\n",
+        "\n",
+        "When deployed on one V100 GPU, the averaged inference time of a request is ~15 seconds."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "bf55e38815dc"
+      },
+      "outputs": [],
+      "source": [
+        "# Set the model_id to a GCS path, like \"gs://GCS_BUCKET/dreambooth/output\", to load the dreambooth finetuned model above.\n",
+        "model, endpoint = deploy_model(\n",
+        "    model_id=\"runwayml/stable-diffusion-v1-5\", task=\"text-to-image\"\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "4ab04da3ec9a"
+      },
+      "outputs": [],
+      "source": [
+        "instances = [\n",
+        "    {\"prompt\": \"a squirrel in Picasso style\"},\n",
+        "    {\"prompt\": \"a dog in Picasso style\"},\n",
+        "    {\"prompt\": \"a cat in Picasso style\"},\n",
+        "    {\"prompt\": \"a deer in Picasso style\"},\n",
+        "]\n",
+        "response = endpoint.predict(instances=instances)\n",
+        "images = [base64_to_image(image) for image in response.predictions]\n",
+        "image_grid(images)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "af21a3cff1e0"
+      },
+      "source": [
+        "Clean up resources:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "911406c1561e"
+      },
+      "outputs": [],
+      "source": [
+        "# Delete models.\n",
+        "model.delete()\n",
+        "\n",
+        "# Undeploy model and delete endpoint.\n",
+        "endpoint.delete(force=True)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "c1e51f764a60"
+      },
+      "source": [
+        "### Text-guided image-to-image"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "fa686a54047c"
+      },
+      "source": [
+        "Deploy the stable diffusion model for the text-guided image-to-image task."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "65e32356fbd1"
+      },
+      "outputs": [],
+      "source": [
+        "# Set the model_id to a GCS path, like \"gs://GCS_BUCKET/dreambooth/output\", to load the dreambooth finetuned model above.\n",
+        "model, endpoint = deploy_model(\n",
+        "    model_id=\"runwayml/stable-diffusion-v1-5\", task=\"image-to-image\"\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "83a50fd4a1ed"
+      },
+      "outputs": [],
+      "source": [
+        "init_image = download_image(\n",
+        "    \"https://raw.githubusercontent.com/CompVis/stable-diffusion/main/assets/stable-samples/img2img/sketch-mountains-input.jpg\"\n",
+        ")\n",
+        "display(init_image)\n",
+        "instances = [\n",
+        "    {\n",
+        "        \"prompt\": \"A fantasy landscape, trending on artstation\",\n",
+        "        \"image\": image_to_base64(init_image),\n",
+        "    },\n",
+        "]\n",
+        "response = endpoint.predict(instances=instances)\n",
+        "images = [base64_to_image(image) for image in response.predictions]\n",
+        "display(images[0])"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "ed3795d474b9"
+      },
+      "source": [
+        "Clean up resources:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b53b883257b4"
+      },
+      "outputs": [],
+      "source": [
+        "# Delete models.\n",
+        "model.delete()\n",
+        "\n",
+        "# Undeploy model and delete endpoint.\n",
+        "endpoint.delete(force=True)"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "name": "model_garden_pytorch_stable_diffusion.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/community/model_garden/model_garden_pytorch_stable_diffusion_inpainting.ipynb

@@ -0,0 +1,577 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "7d9bbf86da5e"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2023 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "1e9c07efb6ac"
+      },
+      "source": [
+        "# Vertex AI Model Garden - Stable Diffusion Inpainting\n",
+        "\n",
+        "<table align=\"left\">\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_stable_diffusion_inpainting.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_stable_diffusion_inpainting.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>                                                                                               <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/notebooks/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_stable_diffusion_inpainting.ipynb\">\n",
+        "      <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "cd8433ec804a"
+      },
+      "source": [
+        "## Overview\n",
+        "\n",
+        "This notebook demonstrates finetuning [runwayml/stable-diffusion-inpainting](https://huggingface.co/runwayml/stable-diffusion-inpainting) with [Dreambooth](https://huggingface.co/docs/diffusers/training/dreambooth) and deploying it on Vertex-AI for online prediction.\n",
+        "\n",
+        "### Objective\n",
+        "\n",
+        "- Finetune the stable-diffusion-inpainting model with [Dreambooth](https://huggingface.co/docs/diffusers/training/dreambooth).\n",
+        "- Upload the model to [Model Registry](https://cloud.google.com/vertex-ai/docs/model-registry/introduction).\n",
+        "- Deploy the model on [Endpoint](https://cloud.google.com/vertex-ai/docs/predictions/using-private-endpoints).\n",
+        "- Run online predictions for image-inpainting.\n",
+        "\n",
+        "### Costs\n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI\n",
+        "* Cloud Storage\n",
+        "\n",
+        "Learn about [Vertex AI pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage pricing](https://cloud.google.com/storage/pricing), and use the [Pricing Calculator](https://cloud.google.com/products/calculator/) to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "264c07757582"
+      },
+      "source": [
+        "## Setup environment\n",
+        "\n",
+        "**NOTE**: Jupyter runs lines prefixed with `!` as shell commands, and it interpolates Python variables prefixed with `$` into these commands."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d73ffa0c0b83"
+      },
+      "source": [
+        "### Colab only"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "2707b02ef5df"
+      },
+      "outputs": [],
+      "source": [
+        "!pip3 install --upgrade google-cloud-aiplatform"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b60a4d7100bf"
+      },
+      "outputs": [],
+      "source": [
+        "from google.colab import auth as google_auth\n",
+        "\n",
+        "google_auth.authenticate_user()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "fb671e75ca7b"
+      },
+      "source": [
+        "### Install dependencies"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "dc8ee367fb42"
+      },
+      "outputs": [],
+      "source": [
+        "# Install gdown for downloading example training images.\n",
+        "!pip install gdown\n",
+        "# Install gsutil for downloading/uploading data from/to Cloud Storage buckets.\n",
+        "!pip install gsutil"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "5244aac3d929"
+      },
+      "source": [
+        "Restart the notebook kernel after installs."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "567212ff53a6"
+      },
+      "outputs": [],
+      "source": [
+        "import IPython\n",
+        "\n",
+        "app = IPython.Application.instance()\n",
+        "app.kernel.do_shutdown(True)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "bb7adab99e41"
+      },
+      "source": [
+        "### Setup Google Cloud project\n",
+        "\n",
+        "1. [Select or create a Google Cloud project](https://console.cloud.google.com/cloud-resource-manager). When you first create an account, you get a $300 free credit towards your compute/storage costs.\n",
+        "\n",
+        "1. [Make sure that billing is enabled for your project](https://cloud.google.com/billing/docs/how-to/modify-project).\n",
+        "\n",
+        "1. [Enable the Vertex AI API and Compute Engine API](https://console.cloud.google.com/flows/enableapi?apiid=aiplatform.googleapis.com,compute_component).\n",
+        "\n",
+        "1. [Create a Cloud Storage bucket](https://cloud.google.com/storage/docs/creating-buckets) for storing experiment outputs."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "6c460088b873"
+      },
+      "source": [
+        "Fill following variables for experiments environment:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "855d6b96f291"
+      },
+      "outputs": [],
+      "source": [
+        "# Cloud project id.\n",
+        "PROJECT_ID = \"\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# The region you want to launch jobs in.\n",
+        "REGION = \"us-central1\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# The Cloud Storage bucket for storing experiments output. Fill it without the 'gs://' prefix.\n",
+        "GCS_BUCKET = \"\"  # @param {type:\"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "e828eb320337"
+      },
+      "source": [
+        "Initialize Vertex-AI API:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "12cd25839741"
+      },
+      "outputs": [],
+      "source": [
+        "from google.cloud import aiplatform\n",
+        "\n",
+        "aiplatform.init(project=PROJECT_ID, location=REGION, staging_bucket=GCS_BUCKET)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "2cc825514deb"
+      },
+      "source": [
+        "### Define constants"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b42bd4fa2b2d"
+      },
+      "outputs": [],
+      "source": [
+        "# The pre-built training docker image. It contains training scripts and models.\n",
+        "TRAIN_DOCKER_URI = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/pytorch-diffusers-train:latest\"\n",
+        "\n",
+        "# The pre-built serving docker image. It contains serving scripts and models.\n",
+        "SERVE_DOCKER_URI = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/pytorch-diffusers-serve:latest\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0c250872074f"
+      },
+      "source": [
+        "### Define common functions"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 2,
+      "metadata": {
+        "id": "8759e624ebc0"
+      },
+      "outputs": [],
+      "source": [
+        "import base64\n",
+        "import os\n",
+        "from datetime import datetime\n",
+        "from io import BytesIO\n",
+        "\n",
+        "import requests\n",
+        "from google.cloud import aiplatform\n",
+        "from PIL import Image\n",
+        "\n",
+        "\n",
+        "def create_job_name(prefix):\n",
+        "    user = os.environ.get(\"USER\")\n",
+        "    now = datetime.now().strftime(\"%Y%m%d_%H%M%S\")\n",
+        "    job_name = f\"{prefix}-{user}-{now}\"\n",
+        "    return job_name\n",
+        "\n",
+        "\n",
+        "def download_image(url):\n",
+        "    response = requests.get(url)\n",
+        "    return Image.open(BytesIO(response.content))\n",
+        "\n",
+        "\n",
+        "def image_to_base64(image, format=\"JPEG\"):\n",
+        "    buffer = BytesIO()\n",
+        "    image.save(buffer, format=format)\n",
+        "    image_str = base64.b64encode(buffer.getvalue()).decode(\"utf-8\")\n",
+        "    return image_str\n",
+        "\n",
+        "\n",
+        "def base64_to_image(image_str):\n",
+        "    image = Image.open(BytesIO(base64.b64decode(image_str)))\n",
+        "    return image\n",
+        "\n",
+        "\n",
+        "def image_grid(imgs, rows=2, cols=2):\n",
+        "    w, h = imgs[0].size\n",
+        "    grid = Image.new(\"RGB\", size=(cols * w, rows * h))\n",
+        "    for i, img in enumerate(imgs):\n",
+        "        grid.paste(img, box=(i % cols * w, i // cols * h))\n",
+        "    return grid\n",
+        "\n",
+        "\n",
+        "def deploy_model(model_id, task):\n",
+        "    model_name = \"stable-diffusion-inpainting\"\n",
+        "    endpoint = aiplatform.Endpoint.create(display_name=f\"{model_name}-endpoint\")\n",
+        "    serving_env = {\n",
+        "        \"MODEL_ID\": model_id,\n",
+        "        \"TASK\": task,\n",
+        "    }\n",
+        "    # If the model_id is a GCS path, use artifact_uri to pass it to serving docker.\n",
+        "    artifact_uri = model_id if model_id.startswith(\"gs://\") else None\n",
+        "    model = aiplatform.Model.upload(\n",
+        "        display_name=model_name,\n",
+        "        serving_container_image_uri=SERVE_DOCKER_URI,\n",
+        "        serving_container_ports=[7080],\n",
+        "        serving_container_predict_route=\"/predictions/diffusers_serving\",\n",
+        "        serving_container_health_route=\"/ping\",\n",
+        "        serving_container_environment_variables=serving_env,\n",
+        "        artifact_uri=artifact_uri,\n",
+        "    )\n",
+        "    model.deploy(\n",
+        "        endpoint=endpoint,\n",
+        "        machine_type=\"n1-standard-8\",\n",
+        "        accelerator_type=\"NVIDIA_TESLA_V100\",\n",
+        "        accelerator_count=1,\n",
+        "        deploy_request_timeout=1800,\n",
+        "    )\n",
+        "    return model, endpoint"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "e70e3519ff8b"
+      },
+      "source": [
+        "## Finetune with Dreambooth"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "f560edbf96c6"
+      },
+      "source": [
+        "This section uses [dreambooth](https://dreambooth.github.io/) to finetune the [stable-diffusion-inpainting](https://huggingface.co/runwayml/stable-diffusion-inpainting) model with [5 dog images](https://drive.google.com/drive/folders/1BO_dyz-p65qhBRRMRA4TbZ8qW4rB99JZ) to personalize the model.\n",
+        "\n",
+        "It finetunes both text encoder and unet of the stable diffusion model up to 800 steps. The whole finetuning job takes 30 minutes to finish using 1 A100 GPU.\n",
+        "\n",
+        "The full model will be saved after the finetuning job finishs and it can be loaded by the [StableDiffusionInpaintPipeline](https://huggingface.co/docs/diffusers/api/pipelines/stable_diffusion/inpaint) to run inference."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "34048707df5c"
+      },
+      "outputs": [],
+      "source": [
+        "# Download example training images.\n",
+        "!gdown --folder https://drive.google.com/drive/folders/1BO_dyz-p65qhBRRMRA4TbZ8qW4rB99JZ\n",
+        "\n",
+        "# Upload data to Cloud Storage bucket.\n",
+        "!gsutil -m cp -r dog/* gs://{GCS_BUCKET}/dreambooth/dog/\n",
+        "!gsutil -m cp -r dog/* gs://{GCS_BUCKET}/dreambooth/dog_class/"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "969cfeb79317"
+      },
+      "source": [
+        "**NOTE**: If the upload step fails due to lacking of permission, you need to [grant the Storage Object Admin role](https://cloud.google.com/storage/docs/access-control/using-iam-permissions) for the Cloud account of the notebook."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "f6d5a05592e1"
+      },
+      "outputs": [],
+      "source": [
+        "# The pre-trained model to be loaded.\n",
+        "model_id = \"runwayml/stable-diffusion-inpainting\"\n",
+        "\n",
+        "# Input and output path.\n",
+        "instance_dir = f\"/gcs/{GCS_BUCKET}/dreambooth/dog\"\n",
+        "class_dir = f\"/gcs/{GCS_BUCKET}/dreambooth/dog_class\"\n",
+        "output_dir = f\"/gcs/{GCS_BUCKET}/dreambooth/output\"\n",
+        "\n",
+        "# Worker pool spec.\n",
+        "machine_type = \"a2-highgpu-1g\"\n",
+        "num_nodes = 1\n",
+        "gpu_type = \"NVIDIA_TESLA_A100\"\n",
+        "num_gpus = 1\n",
+        "\n",
+        "# Setup training job.\n",
+        "job_name = create_job_name(\"dreambooth-stable-diffusion-inpainting\")\n",
+        "job = aiplatform.CustomContainerTrainingJob(\n",
+        "    display_name=job_name,\n",
+        "    container_uri=TRAIN_DOCKER_URI,\n",
+        ")\n",
+        "\n",
+        "# Pass training arguments and launch job.\n",
+        "# See https://github.com/huggingface/diffusers/blob/v0.14.0/examples/research_projects/dreambooth_inpaint/train_dreambooth_inpaint.py#L83\n",
+        "# for a full list of training arguments.\n",
+        "model = job.run(\n",
+        "    args=[\n",
+        "        \"research_projects/dreambooth_inpaint/train_dreambooth_inpaint.py\",\n",
+        "        f\"--pretrained_model_name_or_path={model_id}\",\n",
+        "        \"--train_text_encoder\",\n",
+        "        f\"--instance_data_dir={instance_dir}\",\n",
+        "        f\"--class_data_dir={class_dir}\",\n",
+        "        f\"--output_dir={output_dir}\",\n",
+        "        \"--with_prior_preservation\",\n",
+        "        \"--prior_loss_weight=1.0\",\n",
+        "        \"--instance_prompt='a photo of sks dog'\",\n",
+        "        \"--class_prompt='a photo of dog'\",\n",
+        "        \"--resolution=512\",\n",
+        "        \"--train_batch_size=1\",\n",
+        "        \"--gradient_checkpointing\",\n",
+        "        \"--learning_rate=2e-6\",\n",
+        "        \"--lr_scheduler=constant\",\n",
+        "        \"--lr_warmup_steps=0\",\n",
+        "        \"--num_class_images=200\",\n",
+        "        \"--max_train_steps=800\",\n",
+        "    ],\n",
+        "    replica_count=num_nodes,\n",
+        "    machine_type=machine_type,\n",
+        "    accelerator_type=gpu_type,\n",
+        "    accelerator_count=num_gpus,\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "90d3c379090e"
+      },
+      "source": [
+        "## Upload and deploy models"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "1cc26e68d7b0"
+      },
+      "source": [
+        "This section uploads the model to Model Registry and deploys it on the Endpoint.\n",
+        "\n",
+        "The model deployment step will take ~15 minutes to complete."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "b8bb7d198315"
+      },
+      "source": [
+        "### Image-inpainting"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "79b66382f849"
+      },
+      "source": [
+        "Deploy the stable diffusion model for the image-inpainting task.\n",
+        "\n",
+        "Once deployed, you can send prompts to the endpoint to generated images.\n",
+        "\n",
+        "When deployed on one V100 GPU, the averaged inference time of a request is ~15 seconds."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "a881564da1d8"
+      },
+      "outputs": [],
+      "source": [
+        "# Set the model_id to a GCS path, like \"gs://GCS_BUCKET/dreambooth/output\", to load the dreambooth finetuned model above.\n",
+        "model, endpoint = deploy_model(\n",
+        "    model_id=\"runwayml/stable-diffusion-inpainting\", task=\"image-inpainting\"\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "ca1761afb66f"
+      },
+      "outputs": [],
+      "source": [
+        "img_url = \"https://raw.githubusercontent.com/CompVis/latent-diffusion/main/data/inpainting_examples/bertrand-gabioud-CpuFzIsHYJ0.png\"\n",
+        "mask_url = \"https://raw.githubusercontent.com/CompVis/latent-diffusion/main/data/inpainting_examples/bertrand-gabioud-CpuFzIsHYJ0_mask.png\"\n",
+        "init_image = download_image(img_url).resize((512, 512))\n",
+        "mask_image = download_image(mask_url).resize((512, 512))\n",
+        "display(init_image)\n",
+        "display(mask_image)\n",
+        "\n",
+        "instances = [\n",
+        "    {\n",
+        "        \"prompt\": \"a tree, high resolution, in front of high buildings\",\n",
+        "        \"image\": image_to_base64(init_image),\n",
+        "        \"mask_image\": image_to_base64(mask_image),\n",
+        "    },\n",
+        "]\n",
+        "response = endpoint.predict(instances=instances)\n",
+        "images = [base64_to_image(image) for image in response.predictions]\n",
+        "display(images[0])"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "f12f8d9c2786"
+      },
+      "source": [
+        "### Clean up resources"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "911406c1561e"
+      },
+      "outputs": [],
+      "source": [
+        "# Delete models.\n",
+        "model.delete()\n",
+        "\n",
+        "# Undeploy model and delete endpoint.\n",
+        "endpoint.delete(force=True)"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "name": "model_garden_pytorch_stable_diffusion_inpainting.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/community/model_garden/model_garden_pytorch_vilt_vqa.ipynb

@@ -0,0 +1,392 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "7d9bbf86da5e"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2023 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "2bd716bf3e39"
+      },
+      "source": [
+        "# Vertex AI Model Garden - ViLT VQA\n",
+        "\n",
+        "<table align=\"left\">\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_vilt_vqa.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_vilt_vqa.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>                                                                                               <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/notebooks/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_vilt_vqa.ipynb\">\n",
+        "      <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d8cd12648da4"
+      },
+      "source": [
+        "## Overview\n",
+        "\n",
+        "This notebook demonstrates deploying the pre-trained [ViLT VQA](https://huggingface.co/dandelin/vilt-b32-finetuned-vqa) model on Vertex AI for online prediction.\n",
+        "\n",
+        "### Objective\n",
+        "\n",
+        "- Upload the model to [Model Registry](https://cloud.google.com/vertex-ai/docs/model-registry/introduction).\n",
+        "- Deploy the model on [Endpoint](https://cloud.google.com/vertex-ai/docs/predictions/using-private-endpoints).\n",
+        "- Run online predictions for image captioning.\n",
+        "\n",
+        "### Costs\n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI\n",
+        "* Cloud Storage\n",
+        "\n",
+        "Learn about [Vertex AI pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage pricing](https://cloud.google.com/storage/pricing), and use the [Pricing Calculator](https://cloud.google.com/products/calculator/) to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "264c07757582"
+      },
+      "source": [
+        "## Setup environment\n",
+        "\n",
+        "**NOTE**: Jupyter runs lines prefixed with `!` as shell commands, and it interpolates Python variables prefixed with `$` into these commands."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d73ffa0c0b83"
+      },
+      "source": [
+        "### Colab only"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "2707b02ef5df"
+      },
+      "outputs": [],
+      "source": [
+        "!pip3 install --upgrade google-cloud-aiplatform"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b60a4d7100bf"
+      },
+      "outputs": [],
+      "source": [
+        "from google.colab import auth as google_auth\n",
+        "\n",
+        "google_auth.authenticate_user()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0f826ff482a2"
+      },
+      "source": [
+        "### Setup Google Cloud project\n",
+        "\n",
+        "1. [Select or create a Google Cloud project](https://console.cloud.google.com/cloud-resource-manager). When you first create an account, you get a $300 free credit towards your compute/storage costs.\n",
+        "\n",
+        "1. [Make sure that billing is enabled for your project](https://cloud.google.com/billing/docs/how-to/modify-project).\n",
+        "\n",
+        "1. [Enable the Vertex AI API and Compute Engine API](https://console.cloud.google.com/flows/enableapi?apiid=aiplatform.googleapis.com,compute_component).\n",
+        "\n",
+        "1. [Create a Cloud Storage bucket](https://cloud.google.com/storage/docs/creating-buckets) for storing experiment outputs."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "8958ebc71868"
+      },
+      "source": [
+        "Fill following variables for experiments environment:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "9db30f827a65"
+      },
+      "outputs": [],
+      "source": [
+        "# Cloud project id.\n",
+        "PROJECT_ID = \"\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# The region you want to launch jobs in.\n",
+        "REGION = \"us-central1\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# The Cloud Storage bucket for storing experiments output. Fill it without the 'gs://' prefix.\n",
+        "GCS_BUCKET = \"\"  # @param {type:\"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "92f16e22c20b"
+      },
+      "source": [
+        "Initialize Vertex AI API:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "1680c257acfb"
+      },
+      "outputs": [],
+      "source": [
+        "from google.cloud import aiplatform\n",
+        "\n",
+        "aiplatform.init(project=PROJECT_ID, location=REGION, staging_bucket=GCS_BUCKET)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "6ca48b699d17"
+      },
+      "source": [
+        "### Define constants"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "de9882ea89ea"
+      },
+      "outputs": [],
+      "source": [
+        "# The pre-built serving docker image. It contains serving scripts and models.\n",
+        "SERVE_DOCKER_URI = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/pytorch-transformers-serve:latest\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "10188266a5cd"
+      },
+      "source": [
+        "### Define common functions"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "cac4478ae098"
+      },
+      "outputs": [],
+      "source": [
+        "import base64\n",
+        "import os\n",
+        "from datetime import datetime\n",
+        "from io import BytesIO\n",
+        "\n",
+        "import requests\n",
+        "from google.cloud import aiplatform\n",
+        "from PIL import Image\n",
+        "\n",
+        "\n",
+        "def create_job_name(prefix):\n",
+        "    user = os.environ.get(\"USER\")\n",
+        "    now = datetime.now().strftime(\"%Y%m%d_%H%M%S\")\n",
+        "    job_name = f\"{prefix}-{user}-{now}\"\n",
+        "    return job_name\n",
+        "\n",
+        "\n",
+        "def download_image(url):\n",
+        "    response = requests.get(url)\n",
+        "    return Image.open(BytesIO(response.content))\n",
+        "\n",
+        "\n",
+        "def image_to_base64(image, format=\"JPEG\"):\n",
+        "    buffer = BytesIO()\n",
+        "    image.save(buffer, format=format)\n",
+        "    image_str = base64.b64encode(buffer.getvalue()).decode(\"utf-8\")\n",
+        "    return image_str\n",
+        "\n",
+        "\n",
+        "def base64_to_image(image_str):\n",
+        "    image = Image.open(BytesIO(base64.b64decode(image_str)))\n",
+        "    return image\n",
+        "\n",
+        "\n",
+        "def image_grid(imgs, rows=2, cols=2):\n",
+        "    w, h = imgs[0].size\n",
+        "    grid = Image.new(\"RGB\", size=(cols * w, rows * h))\n",
+        "    for i, img in enumerate(imgs):\n",
+        "        grid.paste(img, box=(i % cols * w, i // cols * h))\n",
+        "    return grid\n",
+        "\n",
+        "\n",
+        "def deploy_model(model_id, task):\n",
+        "    model_name = \"vilt-vqa\"\n",
+        "    endpoint = aiplatform.Endpoint.create(display_name=f\"{model_name}-endpoint\")\n",
+        "    serving_env = {\n",
+        "        \"MODEL_ID\": model_id,\n",
+        "        \"TASK\": task,\n",
+        "    }\n",
+        "    # If the model_id is a GCS path, use artifact_uri to pass it to serving docker.\n",
+        "    artifact_uri = model_id if model_id.startswith(\"gs://\") else None\n",
+        "    model = aiplatform.Model.upload(\n",
+        "        display_name=model_name,\n",
+        "        serving_container_image_uri=SERVE_DOCKER_URI,\n",
+        "        serving_container_ports=[7080],\n",
+        "        serving_container_predict_route=\"/predictions/diffusers_serving\",\n",
+        "        serving_container_health_route=\"/ping\",\n",
+        "        serving_container_environment_variables=serving_env,\n",
+        "        artifact_uri=artifact_uri,\n",
+        "    )\n",
+        "    model.deploy(\n",
+        "        endpoint=endpoint,\n",
+        "        machine_type=\"n1-standard-8\",\n",
+        "        accelerator_type=\"NVIDIA_TESLA_T4\",\n",
+        "        accelerator_count=1,\n",
+        "        deploy_request_timeout=1800,\n",
+        "    )\n",
+        "    return model, endpoint"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d2d72ecdb8c9"
+      },
+      "source": [
+        "## Upload and deploy models"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "9448c5f545fa"
+      },
+      "source": [
+        "This section uploads the pre-trained model to Model Registry and deploys it on the Endpoint with 1 T4 GPU.\n",
+        "\n",
+        "The model deployment step will take ~15 minutes to complete.\n",
+        "\n",
+        "Once deployed, you can send images and questions to get answers."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b4b46c28d8b1"
+      },
+      "outputs": [],
+      "source": [
+        "model, endpoint = deploy_model(\n",
+        "    model_id=\"dandelin/vilt-b32-finetuned-vqa\", task=\"visual-question-answering\"\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "6be655247cb1"
+      },
+      "outputs": [],
+      "source": [
+        "image = download_image(\"http://images.cocodataset.org/val2017/000000039769.jpg\")\n",
+        "display(image)\n",
+        "\n",
+        "question = \"Which cat is bigger?\"\n",
+        "instances = [\n",
+        "    {\"image\": image_to_base64(image), \"text\": question},\n",
+        "]\n",
+        "preds = endpoint.predict(instances=instances).predictions\n",
+        "print(question)\n",
+        "print(preds)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "db7ffebdb4be"
+      },
+      "source": [
+        "### Clean up resources"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "2ccf3714dbe9"
+      },
+      "outputs": [],
+      "source": [
+        "# Delete models.\n",
+        "model.delete()\n",
+        "\n",
+        "# Undeploy model and delete endpoint.\n",
+        "endpoint.delete(force=True)"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "name": "model_garden_pytorch_vilt_vqa.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/community/model_garden/model_garden_pytorch_vit_gpt2_image_captioning.ipynb

@@ -0,0 +1,390 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "7d9bbf86da5e"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2023 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "2bd716bf3e39"
+      },
+      "source": [
+        "# Vertex AI Model Garden - ViT-GPT2 Image Captioning\n",
+        "\n",
+        "<table align=\"left\">\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_vit_gpt2_image_captioning.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_vit_gpt2_image_captioning.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>                                                                                               <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/notebooks/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_vit_gpt2_image_captioning.ipynb\">\n",
+        "      <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d8cd12648da4"
+      },
+      "source": [
+        "## Overview\n",
+        "\n",
+        "This notebook demonstrates deploying the pre-trained [ViT-GPT2 Image Captioning](https://huggingface.co/nlpconnect/vit-gpt2-image-captioning) model on Vertex AI for online prediction.\n",
+        "\n",
+        "### Objective\n",
+        "\n",
+        "- Upload the model to [Model Registry](https://cloud.google.com/vertex-ai/docs/model-registry/introduction).\n",
+        "- Deploy the model on [Endpoint](https://cloud.google.com/vertex-ai/docs/predictions/using-private-endpoints).\n",
+        "- Run online predictions for image captioning.\n",
+        "\n",
+        "### Costs\n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI\n",
+        "* Cloud Storage\n",
+        "\n",
+        "Learn about [Vertex AI pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage pricing](https://cloud.google.com/storage/pricing), and use the [Pricing Calculator](https://cloud.google.com/products/calculator/) to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "264c07757582"
+      },
+      "source": [
+        "## Setup environment\n",
+        "\n",
+        "**NOTE**: Jupyter runs lines prefixed with `!` as shell commands, and it interpolates Python variables prefixed with `$` into these commands."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d73ffa0c0b83"
+      },
+      "source": [
+        "### Colab only"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "2707b02ef5df"
+      },
+      "outputs": [],
+      "source": [
+        "!pip3 install --upgrade google-cloud-aiplatform"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b60a4d7100bf"
+      },
+      "outputs": [],
+      "source": [
+        "from google.colab import auth as google_auth\n",
+        "\n",
+        "google_auth.authenticate_user()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0f826ff482a2"
+      },
+      "source": [
+        "### Setup Google Cloud project\n",
+        "\n",
+        "1. [Select or create a Google Cloud project](https://console.cloud.google.com/cloud-resource-manager). When you first create an account, you get a $300 free credit towards your compute/storage costs.\n",
+        "\n",
+        "1. [Make sure that billing is enabled for your project](https://cloud.google.com/billing/docs/how-to/modify-project).\n",
+        "\n",
+        "1. [Enable the Vertex AI API and Compute Engine API](https://console.cloud.google.com/flows/enableapi?apiid=aiplatform.googleapis.com,compute_component).\n",
+        "\n",
+        "1. [Create a Cloud Storage bucket](https://cloud.google.com/storage/docs/creating-buckets) for storing experiment outputs."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "8958ebc71868"
+      },
+      "source": [
+        "Fill following variables for experiments environment:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "9db30f827a65"
+      },
+      "outputs": [],
+      "source": [
+        "# Cloud project id.\n",
+        "PROJECT_ID = \"\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# The region you want to launch jobs in.\n",
+        "REGION = \"us-central1\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# The Cloud Storage bucket for storing experiments output. Fill it without the 'gs://' prefix.\n",
+        "GCS_BUCKET = \"\"  # @param {type:\"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "92f16e22c20b"
+      },
+      "source": [
+        "Initialize Vertex AI API:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "1680c257acfb"
+      },
+      "outputs": [],
+      "source": [
+        "from google.cloud import aiplatform\n",
+        "\n",
+        "aiplatform.init(project=PROJECT_ID, location=REGION, staging_bucket=GCS_BUCKET)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "6ca48b699d17"
+      },
+      "source": [
+        "### Define constants"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "de9882ea89ea"
+      },
+      "outputs": [],
+      "source": [
+        "# The pre-built serving docker image. It contains serving scripts and models.\n",
+        "SERVE_DOCKER_URI = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/pytorch-transformers-serve:latest\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "10188266a5cd"
+      },
+      "source": [
+        "### Define common functions"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "cac4478ae098"
+      },
+      "outputs": [],
+      "source": [
+        "import base64\n",
+        "import os\n",
+        "from datetime import datetime\n",
+        "from io import BytesIO\n",
+        "\n",
+        "import requests\n",
+        "from google.cloud import aiplatform\n",
+        "from PIL import Image\n",
+        "\n",
+        "\n",
+        "def create_job_name(prefix):\n",
+        "    user = os.environ.get(\"USER\")\n",
+        "    now = datetime.now().strftime(\"%Y%m%d_%H%M%S\")\n",
+        "    job_name = f\"{prefix}-{user}-{now}\"\n",
+        "    return job_name\n",
+        "\n",
+        "\n",
+        "def download_image(url):\n",
+        "    response = requests.get(url)\n",
+        "    return Image.open(BytesIO(response.content))\n",
+        "\n",
+        "\n",
+        "def image_to_base64(image, format=\"JPEG\"):\n",
+        "    buffer = BytesIO()\n",
+        "    image.save(buffer, format=format)\n",
+        "    image_str = base64.b64encode(buffer.getvalue()).decode(\"utf-8\")\n",
+        "    return image_str\n",
+        "\n",
+        "\n",
+        "def base64_to_image(image_str):\n",
+        "    image = Image.open(BytesIO(base64.b64decode(image_str)))\n",
+        "    return image\n",
+        "\n",
+        "\n",
+        "def image_grid(imgs, rows=2, cols=2):\n",
+        "    w, h = imgs[0].size\n",
+        "    grid = Image.new(\"RGB\", size=(cols * w, rows * h))\n",
+        "    for i, img in enumerate(imgs):\n",
+        "        grid.paste(img, box=(i % cols * w, i // cols * h))\n",
+        "    return grid\n",
+        "\n",
+        "\n",
+        "def deploy_model(model_id, task):\n",
+        "    model_name = \"vit-gpt2-image-captioning\"\n",
+        "    endpoint = aiplatform.Endpoint.create(display_name=f\"{model_name}-endpoint\")\n",
+        "    serving_env = {\n",
+        "        \"MODEL_ID\": model_id,\n",
+        "        \"TASK\": task,\n",
+        "    }\n",
+        "    # If the model_id is a GCS path, use artifact_uri to pass it to serving docker.\n",
+        "    artifact_uri = model_id if model_id.startswith(\"gs://\") else None\n",
+        "    model = aiplatform.Model.upload(\n",
+        "        display_name=model_name,\n",
+        "        serving_container_image_uri=SERVE_DOCKER_URI,\n",
+        "        serving_container_ports=[7080],\n",
+        "        serving_container_predict_route=\"/predictions/diffusers_serving\",\n",
+        "        serving_container_health_route=\"/ping\",\n",
+        "        serving_container_environment_variables=serving_env,\n",
+        "        artifact_uri=artifact_uri,\n",
+        "    )\n",
+        "    model.deploy(\n",
+        "        endpoint=endpoint,\n",
+        "        machine_type=\"n1-standard-8\",\n",
+        "        accelerator_type=\"NVIDIA_TESLA_T4\",\n",
+        "        accelerator_count=1,\n",
+        "        deploy_request_timeout=1800,\n",
+        "    )\n",
+        "    return model, endpoint"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d2d72ecdb8c9"
+      },
+      "source": [
+        "## Upload and deploy models"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "9448c5f545fa"
+      },
+      "source": [
+        "This section uploads the pre-trained model to Model Registry and deploys it on the Endpoint with 1 T4 GPU.\n",
+        "\n",
+        "The model deployment step will take ~15 minutes to complete.\n",
+        "\n",
+        "Once deployed, you can send images to get descriptions."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b4b46c28d8b1"
+      },
+      "outputs": [],
+      "source": [
+        "model, endpoint = deploy_model(\n",
+        "    model_id=\"nlpconnect/vit-gpt2-image-captioning\", task=\"image-to-text\"\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "6be655247cb1"
+      },
+      "outputs": [],
+      "source": [
+        "image = download_image(\"http://images.cocodataset.org/val2017/000000039769.jpg\")\n",
+        "display(image)\n",
+        "\n",
+        "instances = [\n",
+        "    {\"image\": image_to_base64(image)},\n",
+        "]\n",
+        "preds = endpoint.predict(instances=instances).predictions\n",
+        "print(preds)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "db7ffebdb4be"
+      },
+      "source": [
+        "### Clean up resources"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "2ccf3714dbe9"
+      },
+      "outputs": [],
+      "source": [
+        "# Delete models.\n",
+        "model.delete()\n",
+        "\n",
+        "# Undeploy model and delete endpoint.\n",
+        "endpoint.delete(force=True)"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "name": "model_garden_pytorch_vit_gpt2_image_captioning.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/community/model_garden/model_garden_tfvision_image_classification.ipynb

@@ -0,0 +1,901 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "ur8xi4C7S06n"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2023 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "TirJ-SGQseby"
+      },
+      "source": [
+        "# Vertex AI Model Garden TFVision With Image Classification\n",
+        "\n",
+        "<table align=\"left\">\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_tfvision_image_classification.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_tfvision_image_classification.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>                                                                                               <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/notebooks/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/main/notebooks/community/model_garden/model_garden_tfvision_image_classification.ipynb\">\n",
+        "      <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "dwGLvtIeECLK"
+      },
+      "source": [
+        "**_NOTE_**: This notebook has been tested in the following environment:\n",
+        "\n",
+        "* Python version = 3.9"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "tvgnzT1CKxrO"
+      },
+      "source": [
+        "## Overview\n",
+        "\n",
+        "This notebook demonstrates how to use [TFVision](https://github.com/tensorflow/models/blob/master/official/vision/MODEL_GARDEN.md) in Vertex AI Model Garden.\n",
+        "\n",
+        "### Objective\n",
+        "\n",
+        "* Train new models\n",
+        "  * Convert input data to training formats\n",
+        "  * Create [hyperparameter tuning jobs](https://cloud.google.com/vertex-ai/docs/training/hyperparameter-tuning-overview) to train new models\n",
+        "  * Find and export best models\n",
+        "\n",
+        "* Test trained models\n",
+        "  * Upload models to model registry\n",
+        "  * Deploy uploaded models\n",
+        "  * Run predictions\n",
+        "\n",
+        "* Cleanup resources\n",
+        "\n",
+        "### Costs\n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI\n",
+        "* Cloud Storage\n",
+        "\n",
+        "Learn about [Vertex AI\n",
+        "pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage\n",
+        "pricing](https://cloud.google.com/storage/pricing), and use the [Pricing\n",
+        "Calculator](https://cloud.google.com/products/calculator/)\n",
+        "to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "KEukV6uRk_S3"
+      },
+      "source": [
+        "## Before you begin"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "z__i0w0lCAsW"
+      },
+      "source": [
+        "### Colab Only\n",
+        "Run the following commands for colab and skip this section if you use workbench."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "Jvqs-ehKlaYh"
+      },
+      "outputs": [],
+      "source": [
+        "if \"google.colab\" in str(get_ipython()):\n",
+        "    ! pip3 install --upgrade google-cloud-aiplatform\n",
+        "\n",
+        "    # Automatically restart kernel after installs\n",
+        "    import IPython\n",
+        "\n",
+        "    app = IPython.Application.instance()\n",
+        "    app.kernel.do_shutdown(True)\n",
+        "\n",
+        "    from google.colab import auth as google_auth\n",
+        "\n",
+        "    google_auth.authenticate_user()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "BF1j6f9HApxa"
+      },
+      "source": [
+        "### Set up your Google Cloud project\n",
+        "\n",
+        "**The following steps are required, regardless of your notebook environment.**\n",
+        "\n",
+        "1. [Select or create a Google Cloud project](https://console.cloud.google.com/cloud-resource-manager). When you first create an account, you get a $300 free credit towards your compute/storage costs.\n",
+        "\n",
+        "1. [Make sure that billing is enabled for your project](https://cloud.google.com/billing/docs/how-to/modify-project).\n",
+        "\n",
+        "1. [Enable the Vertex AI API and Compute Engine API](https://console.cloud.google.com/flows/enableapi?apiid=aiplatform.googleapis.com,compute_component).\n",
+        "1. If you are running this notebook locally, you will need to install the [Cloud SDK](https://cloud.google.com/sdk).\n",
+        "\n",
+        "1. Enter your project ID in the cell below. Then run the cell to make sure the\n",
+        "Cloud SDK uses the right project for all the commands in this notebook.\n",
+        "\n",
+        "**Note**: Jupyter runs lines prefixed with `!` as shell commands, and it interpolates Python variables prefixed with `$` into these commands."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "9wExiMUxFk91"
+      },
+      "outputs": [],
+      "source": [
+        "import os\n",
+        "\n",
+        "from google.cloud import aiplatform\n",
+        "\n",
+        "# The project and bucket are for experiments below.\n",
+        "PROJECT_ID = \"\"  # @param {type:\"string\"}\n",
+        "BUCKET_URI = \"\"  # @param {type:\"string\"}\n",
+        "REGION = \"us-central1\"\n",
+        "\n",
+        "! gcloud config set project $PROJECT_ID\n",
+        "\n",
+        "STAGING_BUCKET = os.path.join(BUCKET_URI, \"temporal\")\n",
+        "CHECKPOINT_BUCKET = os.path.join(BUCKET_URI, \"ckpt\")\n",
+        "\n",
+        "aiplatform.init(project=PROJECT_ID, location=REGION, staging_bucket=STAGING_BUCKET)\n",
+        "\n",
+        "# Download config files.\n",
+        "CONFIG_DIR = os.path.join(BUCKET_URI, \"config\")\n",
+        "\n",
+        "\n",
+        "def upload_config_to_gcs(url):\n",
+        "    filename = os.path.basename(url)\n",
+        "    destination = os.path.join(CONFIG_DIR, filename)\n",
+        "    print(\"Copy\", url, \"to\", destination)\n",
+        "    ! wget \"$url\" -O \"$filename\"\n",
+        "    ! gsutil cp \"$filename\" \"$destination\"\n",
+        "\n",
+        "\n",
+        "upload_config_to_gcs(\n",
+        "    \"https://raw.githubusercontent.com/tensorflow/models/master/official/vision/configs/experiments/image_classification/imagenet_resnet50_gpu.yaml\"\n",
+        ")\n",
+        "upload_config_to_gcs(\n",
+        "    \"https://raw.githubusercontent.com/tensorflow/models/master/official/vision/configs/experiments/image_classification/imagenet_resnetrs50_i160_gpu.yaml\"\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "n6IFz75WGCam"
+      },
+      "source": [
+        "### Define constants"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "riG_qUokg0XZ"
+      },
+      "outputs": [],
+      "source": [
+        "OBJECTIVE = \"icn\"\n",
+        "\n",
+        "# Data converter constants.\n",
+        "DATA_CONVERTER_JOB_PREFIX = \"data_converter\"\n",
+        "DATA_CONVERTER_CONTAINER = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/data-converter:latest\"\n",
+        "DATA_CONVERTER_MACHINE_TYPE = \"n1-highmem-8\"\n",
+        "\n",
+        "\n",
+        "# Training constants.\n",
+        "TRAINING_JOB_PREFIX = \"train\"\n",
+        "TRAIN_CONTAINER_URI = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/tfvision-oss:latest\"\n",
+        "TRAIN_MACHINE_TYPE = \"n1-highmem-16\"\n",
+        "TRAIN_ACCELERATOR_TYPE = \"NVIDIA_TESLA_P100\"\n",
+        "TRAIN_NUM_GPU = 1\n",
+        "\n",
+        "# Evaluation constants.\n",
+        "EVALUATION_METRIC = \"accuracy\"\n",
+        "\n",
+        "# Export constants.\n",
+        "EXPORT_JOB_PREFIX = \"export\"\n",
+        "EXPORT_CONTAINER_URI = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/tfvision-serving:latest\"\n",
+        "EXPORT_MACHINE_TYPE = \"n1-highmem-8\"\n",
+        "\n",
+        "# Prediction constants.\n",
+        "# You can deploy models with\n",
+        "#   pre-build-dockers: https://cloud.google.com/vertex-ai/docs/predictions/pre-built-containers.\n",
+        "#   and optimized tensorflow runtime dockers: https://cloud.google.com/vertex-ai/docs/predictions/optimized-tensorflow-runtime.\n",
+        "# The example in this notebook uses optimized tensorflow runtime dockers.\n",
+        "# You can adjust accelerator types and machine types to get faster predictions.\n",
+        "PREDICTION_CONTAINER_URI = (\n",
+        "    \"us-docker.pkg.dev/vertex-ai-restricted/prediction/tf_opt-gpu.2-11:latest\"\n",
+        ")\n",
+        "SERVING_CONTAINER_ARGS = [\"--allow_precompilation\", \"--allow_compression\"]\n",
+        "PREDICTION_ACCELERATOR_TYPE = \"NVIDIA_TESLA_T4\"\n",
+        "PREDICTION_MACHINE_TYPE = \"n1-standard-4\"\n",
+        "UPLOAD_JOB_PREFIX = \"upload\"\n",
+        "DEPLOY_JOB_PREFIX = \"deploy\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "ZZFPe_GezXg8"
+      },
+      "source": [
+        "### Define common libraries"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "XcYUGwr-AJGY"
+      },
+      "outputs": [],
+      "source": [
+        "import base64\n",
+        "import json\n",
+        "from datetime import datetime\n",
+        "from io import BytesIO\n",
+        "from typing import Dict, List, Union\n",
+        "\n",
+        "import matplotlib.pyplot as plt\n",
+        "import numpy\n",
+        "import tensorflow as tf\n",
+        "import yaml\n",
+        "from google.protobuf import json_format\n",
+        "from google.protobuf.struct_pb2 import Value\n",
+        "from PIL import Image\n",
+        "\n",
+        "\n",
+        "def get_job_name_with_datetime(prefix: str):\n",
+        "    return prefix + datetime.now().strftime(\"_%Y%m%d_%H%M%S\")\n",
+        "\n",
+        "\n",
+        "def predict_custom_trained_model(\n",
+        "    project: str,\n",
+        "    endpoint_id: str,\n",
+        "    instances: Union[Dict, List[Dict]],\n",
+        "    location: str = \"us-central1\",\n",
+        "    api_endpoint: str = \"us-central1-aiplatform.googleapis.com\",\n",
+        "):\n",
+        "    # The AI Platform services require regional API endpoints.\n",
+        "    client_options = {\"api_endpoint\": api_endpoint}\n",
+        "    # Initialize client that will be used to create and send requests.\n",
+        "    # This client only needs to be created once, and can be reused for multiple requests.\n",
+        "    client = aiplatform.gapic.PredictionServiceClient(client_options=client_options)\n",
+        "    parameters_dict = {}\n",
+        "    parameters = json_format.ParseDict(parameters_dict, Value())\n",
+        "    endpoint = client.endpoint_path(\n",
+        "        project=project, location=location, endpoint=endpoint_id\n",
+        "    )\n",
+        "    response = client.predict(\n",
+        "        endpoint=endpoint, instances=instances, parameters=parameters\n",
+        "    )\n",
+        "    return response.predictions, response.deployed_model_id\n",
+        "\n",
+        "\n",
+        "def load_img(path):\n",
+        "    img = tf.io.read_file(path)\n",
+        "    img = tf.image.decode_jpeg(img, channels=3)\n",
+        "    return Image.fromarray(numpy.uint8(img)).convert(\"RGB\")\n",
+        "\n",
+        "\n",
+        "def display_image(image):\n",
+        "    _ = plt.figure(figsize=(20, 15))\n",
+        "    plt.grid(False)\n",
+        "    plt.imshow(image)\n",
+        "\n",
+        "\n",
+        "def get_prediction_instances(test_filepath, new_width=-1):\n",
+        "    if new_width <= 0:\n",
+        "        with tf.io.gfile.GFile(test_filepath, \"rb\") as input_file:\n",
+        "            encoded_string = base64.b64encode(input_file.read()).decode(\"utf-8\")\n",
+        "    else:\n",
+        "        img = load_img(test_filepath)\n",
+        "        width, height = img.size\n",
+        "        print(\"original input image size: \", width, \" , \", height)\n",
+        "        new_height = int(height * new_width / width)\n",
+        "        new_img = img.resize((new_width, new_height))\n",
+        "        print(\"resized input image size: \", new_width, \" , \", new_height)\n",
+        "        buffered = BytesIO()\n",
+        "        new_img.save(buffered, format=\"JPEG\")\n",
+        "        encoded_string = base64.b64encode(buffered.getvalue()).decode(\"utf-8\")\n",
+        "\n",
+        "    instances = [\n",
+        "        {\n",
+        "            \"encoded_image\": {\"b64\": encoded_string},\n",
+        "        }\n",
+        "    ]\n",
+        "    return instances\n",
+        "\n",
+        "\n",
+        "def get_label_map(label_map_yaml_filepath):\n",
+        "    with tf.io.gfile.GFile(label_map_yaml_filepath, \"rb\") as input_file:\n",
+        "        label_map = yaml.safe_load(input_file.read())\n",
+        "    return label_map\n",
+        "\n",
+        "\n",
+        "def get_best_trial(model_dir, max_trial_count, evaluation_metric):\n",
+        "    best_trial_dir = \"\"\n",
+        "    best_trial_evaluation_results = {}\n",
+        "    best_performance = -1\n",
+        "\n",
+        "    for i in range(max_trial_count):\n",
+        "        current_trial = i + 1\n",
+        "        current_trial_dir = os.path.join(model_dir, \"trial_\" + str(current_trial))\n",
+        "        current_trial_best_ckpt_dir = os.path.join(current_trial_dir, \"best_ckpt\")\n",
+        "        current_trial_best_ckpt_evaluation_filepath = os.path.join(\n",
+        "            current_trial_best_ckpt_dir, \"info.json\"\n",
+        "        )\n",
+        "        with tf.io.gfile.GFile(current_trial_best_ckpt_evaluation_filepath, \"rb\") as f:\n",
+        "            eval_metric_results = json.load(f)\n",
+        "            current_performance = eval_metric_results[evaluation_metric]\n",
+        "            if current_performance > best_performance:\n",
+        "                best_performance = current_performance\n",
+        "                best_trial_dir = current_trial_dir\n",
+        "                best_trial_evaluation_results = eval_metric_results\n",
+        "    return best_trial_dir, best_trial_evaluation_results\n",
+        "\n",
+        "\n",
+        "def upload_checkpoint_to_gcs(checkpoint_url):\n",
+        "    filename = os.path.basename(checkpoint_url)\n",
+        "    checkpoint_name = filename.replace(\".tar.gz\", \"\")\n",
+        "    print(\"Download checkpoint from\", checkpoint_url, \"and store to\", CHECKPOINT_BUCKET)\n",
+        "    ! wget $checkpoint_url -O $filename\n",
+        "    ! mkdir -p $checkpoint_name\n",
+        "    ! tar -xvzf $filename -C $checkpoint_name\n",
+        "\n",
+        "    # Search for relative path to the checkpoint.\n",
+        "    checkpoint_path = None\n",
+        "    for root, dirs, files in os.walk(checkpoint_name):\n",
+        "        for file in files:\n",
+        "            if file.endswith(\".index\"):\n",
+        "                checkpoint_path = os.path.join(root, os.path.splitext(file)[0])\n",
+        "                checkpoint_path = os.path.relpath(checkpoint_path, checkpoint_name)\n",
+        "                break\n",
+        "\n",
+        "    ! gsutil cp -r $checkpoint_name $CHECKPOINT_BUCKET\n",
+        "    checkpoint_uri = os.path.join(CHECKPOINT_BUCKET, checkpoint_name, checkpoint_path)\n",
+        "    print(\"Checkpoint uploaded to\", checkpoint_uri)\n",
+        "    return checkpoint_uri"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "RB_xY9ipr7ZU"
+      },
+      "source": [
+        "## Train new models\n",
+        "This section shows how to train new models.\n",
+        "1. Convert input data to training formats\n",
+        "2. Create hyperparameter tuning jobs to train new models\n",
+        "3. Find and export best models\n",
+        "\n",
+        "If you already trained models, please go to the section `Test Trained models`."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "zgPO1eR3CYjk"
+      },
+      "source": [
+        "### Prepare input data for training\n",
+        "\n",
+        "Prepare data in the format as described [here](https://cloud.google.com/vertex-ai/docs/image-data/classification/prepare-data), and then convert them to the training formats as below:\n",
+        "\n",
+        "* `input_file_path`: The input file path for preparing data.\n",
+        "* `input_file_type`: The input file type, such as csv or jsonl.\n",
+        "* `split_ratio`: The proportion of data to split into train/validation/test.\n",
+        "* `num_shard`: The number of shards for train/validation/test.\n",
+        "* `output_dir`: The output directory, which will container prepared train/test/validation data."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "IndQ_m6ddUEM"
+      },
+      "outputs": [],
+      "source": [
+        "# This job will convert input data as training format, with given split ratios\n",
+        "# and number of shards on train/test/validation.\n",
+        "\n",
+        "data_converter_job_name = get_job_name_with_datetime(\n",
+        "    DATA_CONVERTER_JOB_PREFIX + \"_\" + OBJECTIVE\n",
+        ")\n",
+        "\n",
+        "input_file_path = \"\"  # @param {type:\"string\"}\n",
+        "input_file_type = \"csv\"  # @param [\"csv\", \"jsonl\"]\n",
+        "num_classes = 5  # @param {type:\"integer\"}\n",
+        "split_ratio = \"0.8,0.1,0.1\"\n",
+        "num_shard = \"10,10,10\"\n",
+        "data_converter_output_dir = os.path.join(BUCKET_URI, data_converter_job_name)\n",
+        "\n",
+        "\n",
+        "worker_pool_specs = [\n",
+        "    {\n",
+        "        \"machine_spec\": {\n",
+        "            \"machine_type\": DATA_CONVERTER_MACHINE_TYPE,\n",
+        "        },\n",
+        "        \"replica_count\": 1,\n",
+        "        \"container_spec\": {\n",
+        "            \"image_uri\": DATA_CONVERTER_CONTAINER,\n",
+        "            \"command\": [],\n",
+        "            \"args\": [\n",
+        "                \"--input_file_path=%s\" % input_file_path,\n",
+        "                \"--input_file_type=%s\" % input_file_type,\n",
+        "                \"--objective=%s\" % OBJECTIVE,\n",
+        "                \"--num_shard=%s\" % num_shard,\n",
+        "                \"--split_ratio=%s\" % split_ratio,\n",
+        "                \"--output_dir=%s\" % data_converter_output_dir,\n",
+        "            ],\n",
+        "        },\n",
+        "    }\n",
+        "]\n",
+        "\n",
+        "data_converter_custom_job = aiplatform.CustomJob(\n",
+        "    display_name=data_converter_job_name,\n",
+        "    project=PROJECT_ID,\n",
+        "    worker_pool_specs=worker_pool_specs,\n",
+        "    staging_bucket=STAGING_BUCKET,\n",
+        ")\n",
+        "\n",
+        "data_converter_custom_job.run()\n",
+        "\n",
+        "input_train_data_path = os.path.join(data_converter_output_dir, \"train.tfrecord*\")\n",
+        "input_validation_data_path = os.path.join(data_converter_output_dir, \"val.tfrecord*\")\n",
+        "label_map_path = os.path.join(data_converter_output_dir, \"label_map.yaml\")\n",
+        "print(\"input_train_data_path for training: \", input_train_data_path)\n",
+        "print(\"input_validation_data_path for training: \", input_validation_data_path)\n",
+        "print(\"label_map_path for prediction: \", label_map_path)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "aaff6f5be7f6"
+      },
+      "source": [
+        "### Create a Vertex AI custom job with hyperparameter tuning\n",
+        "\n",
+        "You use the Vertex AI SDK to create and run the hyperparameter tuning job with Vertex AI Model Garden Training Dockers.\n",
+        "\n",
+        "#### Define the following specifications\n",
+        "* `worker_pool_specs`: Dictionary specifying the machine type and Docker image. This example defines a single node cluster with one `n1-standard-4` machine with two `NVIDIA_TESLA_T4` GPUs.\n",
+        "* `parameter_spec`: Dictionary specifying the parameters to optimize. The dictionary key is the string assigned to the command line argument for each hyperparameter in your training application code, and the dictionary value is the parameter specification. The parameter specification includes the type, min/max values, and scale for the hyperparameter.\n",
+        "* `metric_spec`: Dictionary specifying the metric to optimize. The dictionary key is the `hyperparameter_metric_tag` that you set in your training application code, and the value is the optimization goal."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "um_XKbmpTaHx"
+      },
+      "outputs": [],
+      "source": [
+        "from google.cloud.aiplatform import hyperparameter_tuning as hpt\n",
+        "\n",
+        "# Input train and validation datasets can be found from the section above\n",
+        "# `Convert input data for training`.\n",
+        "# Set prepared datasets if exists.\n",
+        "# input_train_data_path = ''\n",
+        "# input_validation_data_path = ''\n",
+        "\n",
+        "experiment = \"ViT-s16\"  # @param [\"ResNet-50\",\"ResNet-RS-50\",\"Efficientnetv2-m\",\"ViT-ti16\",\"ViT-s16\",\"ViT-b16\",\"ViT-l16\"]\n",
+        "\n",
+        "train_job_name = get_job_name_with_datetime(TRAINING_JOB_PREFIX + \"_\" + OBJECTIVE)\n",
+        "model_dir = os.path.join(BUCKET_URI, train_job_name)\n",
+        "\n",
+        "# The arguments here are mainly for test purposes. Please update them\n",
+        "# to get better performances.\n",
+        "common_args = {\n",
+        "    \"input_train_data_path\": input_train_data_path,\n",
+        "    \"input_validation_data_path\": input_validation_data_path,\n",
+        "    \"objective\": OBJECTIVE,\n",
+        "    \"model_dir\": model_dir,\n",
+        "    \"num_classes\": num_classes,\n",
+        "    \"global_batch_size\": 4,\n",
+        "    \"prefetch_buffer_size\": 32,\n",
+        "    \"train_steps\": 2000,\n",
+        "    \"input_size\": \"224,224\",\n",
+        "}\n",
+        "\n",
+        "# Arguments for different experiments.\n",
+        "experiment_container_args_dict = {\n",
+        "    \"ResNet-50\": dict(\n",
+        "        common_args,\n",
+        "        **{\n",
+        "            \"experiment\": \"resnet_imagenet\",\n",
+        "            \"config_file\": os.path.join(CONFIG_DIR, \"imagenet_resnet50_gpu.yaml\"),\n",
+        "        },\n",
+        "    ),\n",
+        "    \"ResNet-RS-50\": dict(\n",
+        "        common_args,\n",
+        "        **{\n",
+        "            \"experiment\": \"resnet_rs_imagenet\",\n",
+        "            \"config_file\": os.path.join(\n",
+        "                CONFIG_DIR, \"imagenet_resnetrs50_i160_gpu.yaml\"\n",
+        "            ),\n",
+        "            \"init_checkpoint\": \"https://storage.googleapis.com/tf_model_garden/vision/resnet-rs/resnet-rs-50-i160.tar.gz\",\n",
+        "            \"input_size\": \"160,160\",\n",
+        "        },\n",
+        "    ),\n",
+        "    \"Efficientnetv2-m\": dict(\n",
+        "        common_args,\n",
+        "        **{\n",
+        "            \"experiment\": \"hub_model\",\n",
+        "        },\n",
+        "    ),\n",
+        "    \"ViT-ti16\": dict(\n",
+        "        common_args,\n",
+        "        **{\n",
+        "            \"experiment\": \"deit_imagenet_pretrain\",\n",
+        "            \"model_name\": \"vit-ti16\",\n",
+        "            \"init_checkpoint\": \"https://storage.googleapis.com/tf_model_garden/vision/vit/vit-deit-imagenet-ti16.tar.gz\",\n",
+        "            \"input_size\": \"224,224\",\n",
+        "        },\n",
+        "    ),\n",
+        "    \"ViT-s16\": dict(\n",
+        "        common_args,\n",
+        "        **{\n",
+        "            \"experiment\": \"deit_imagenet_pretrain\",\n",
+        "            \"model_name\": \"vit-s16\",\n",
+        "            \"init_checkpoint\": \"https://storage.googleapis.com/tf_model_garden/vision/vit/vit-deit-imagenet-s16.tar.gz\",\n",
+        "            \"input_size\": \"224,224\",\n",
+        "        },\n",
+        "    ),\n",
+        "    \"ViT-b16\": dict(\n",
+        "        common_args,\n",
+        "        **{\n",
+        "            \"experiment\": \"deit_imagenet_pretrain\",\n",
+        "            \"model_name\": \"vit-b16\",\n",
+        "            \"init_checkpoint\": \"https://storage.googleapis.com/tf_model_garden/vision/vit/vit-deit-imagenet-b16.tar.gz\",\n",
+        "            \"input_size\": \"224,224\",\n",
+        "        },\n",
+        "    ),\n",
+        "    \"ViT-l16\": dict(\n",
+        "        common_args,\n",
+        "        **{\n",
+        "            \"experiment\": \"deit_imagenet_pretrain\",\n",
+        "            \"model_name\": \"vit-l16\",\n",
+        "            \"init_checkpoint\": \"https://storage.googleapis.com/tf_model_garden/vision/vit/vit-deit-imagenet-l16.tar.gz\",\n",
+        "            \"input_size\": \"224,224\",\n",
+        "        },\n",
+        "    ),\n",
+        "}\n",
+        "experiment_container_args = experiment_container_args_dict[experiment]\n",
+        "\n",
+        "# Copy checkpoint to GCS bucket if specified.\n",
+        "init_checkpoint = experiment_container_args.get(\"init_checkpoint\")\n",
+        "if init_checkpoint:\n",
+        "    experiment_container_args[\"init_checkpoint\"] = upload_checkpoint_to_gcs(\n",
+        "        init_checkpoint\n",
+        "    )\n",
+        "\n",
+        "worker_pool_specs = [\n",
+        "    {\n",
+        "        \"machine_spec\": {\n",
+        "            \"machine_type\": TRAIN_MACHINE_TYPE,\n",
+        "            \"accelerator_type\": TRAIN_ACCELERATOR_TYPE,\n",
+        "            # Each training job uses TRAIN_NUM_GPU GPUs.\n",
+        "            \"accelerator_count\": TRAIN_NUM_GPU,\n",
+        "        },\n",
+        "        \"replica_count\": 1,\n",
+        "        \"container_spec\": {\n",
+        "            \"image_uri\": TRAIN_CONTAINER_URI,\n",
+        "            \"args\": [\n",
+        "                \"--mode=train_and_eval\",\n",
+        "                \"--params_override=runtime.num_gpus=%d\" % TRAIN_NUM_GPU,\n",
+        "            ]\n",
+        "            + [\"--{}={}\".format(k, v) for k, v in experiment_container_args.items()],\n",
+        "        },\n",
+        "    }\n",
+        "]\n",
+        "\n",
+        "metric_spec = {\"model_performance\": \"maximize\"}\n",
+        "\n",
+        "\n",
+        "LEARNING_RATES = [5e-4, 1e-3]\n",
+        "# Models will be trained with each learning rate separately and max trial count is the number of learning rates.\n",
+        "MAX_TRIAL_COUNT = len(LEARNING_RATES)\n",
+        "parameter_spec = {\n",
+        "    \"learning_rate\": hpt.DiscreteParameterSpec(values=LEARNING_RATES, scale=\"linear\"),\n",
+        "}\n",
+        "\n",
+        "print(worker_pool_specs, metric_spec, parameter_spec)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "HwcCjwlBTQIz"
+      },
+      "source": [
+        "#### Run the hyperparameter tuning job\n",
+        "* `max_trial_count`: Sets an upper bound on the number of trials the service will run. The recommended practice is to start with a smaller number of trials and get a sense of how impactful your chosen hyperparameters are before scaling up.\n",
+        "\n",
+        "* `parallel_trial_count`:  If you use parallel trials, the service provisions multiple training processing clusters. The worker pool spec that you specify when creating the job is used for each individual training cluster.  Increasing the number of parallel trials reduces the amount of time the hyperparameter tuning job takes to run; however, it can reduce the effectiveness of the job overall. This is because the default tuning strategy uses results of previous trials to inform the assignment of values in subsequent trials.\n",
+        "\n",
+        "* `search_algorithm`: The available search algorithms are grid, random, or default (None). The default option applies Bayesian optimization to search the space of possible hyperparameter values and is the recommended algorithm.\n",
+        "\n",
+        "Click on the generated link in the output to see your run in the Cloud Console."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "aec22792ee84"
+      },
+      "outputs": [],
+      "source": [
+        "train_custom_job = aiplatform.CustomJob(\n",
+        "    display_name=train_job_name,\n",
+        "    project=PROJECT_ID,\n",
+        "    worker_pool_specs=worker_pool_specs,\n",
+        "    staging_bucket=STAGING_BUCKET,\n",
+        ")\n",
+        "\n",
+        "train_hpt_job = aiplatform.HyperparameterTuningJob(\n",
+        "    display_name=train_job_name,\n",
+        "    custom_job=train_custom_job,\n",
+        "    metric_spec=metric_spec,\n",
+        "    parameter_spec=parameter_spec,\n",
+        "    max_trial_count=MAX_TRIAL_COUNT,\n",
+        "    parallel_trial_count=MAX_TRIAL_COUNT,\n",
+        "    project=PROJECT_ID,\n",
+        "    search_algorithm=None,\n",
+        ")\n",
+        "\n",
+        "train_hpt_job.run()\n",
+        "\n",
+        "print(\"experiment is: \", experiment)\n",
+        "print(\"model_dir is: \", model_dir)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "mV-Djz-frBni"
+      },
+      "source": [
+        "### Export best models as TF Saved Model format"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "09Rz1AYspK19"
+      },
+      "outputs": [],
+      "source": [
+        "# This job will export models from TF checkpoints to TF saved model format.\n",
+        "# model_dir is from the section above.\n",
+        "best_trial_dir, best_trial_evaluation_results = get_best_trial(\n",
+        "    model_dir, MAX_TRIAL_COUNT, EVALUATION_METRIC\n",
+        ")\n",
+        "print(\"best_trial_dir: \", best_trial_dir)\n",
+        "print(\"best_trial_evaluation_results: \", best_trial_evaluation_results)\n",
+        "\n",
+        "worker_pool_specs = [\n",
+        "    {\n",
+        "        \"machine_spec\": {\n",
+        "            \"machine_type\": EXPORT_MACHINE_TYPE,\n",
+        "        },\n",
+        "        \"replica_count\": 1,\n",
+        "        \"container_spec\": {\n",
+        "            \"image_uri\": EXPORT_CONTAINER_URI,\n",
+        "            \"command\": [],\n",
+        "            \"args\": [\n",
+        "                \"--objective=%s\" % OBJECTIVE,\n",
+        "                \"--input_image_size=%s\" % experiment_container_args[\"input_size\"],\n",
+        "                \"--experiment=%s\" % experiment_container_args[\"experiment\"],\n",
+        "                \"--config_file=%s/params.yaml\" % best_trial_dir,\n",
+        "                \"--checkpoint_path=%s/best_ckpt\" % best_trial_dir,\n",
+        "                \"--export_dir=%s/best_model\" % model_dir,\n",
+        "            ],\n",
+        "        },\n",
+        "    }\n",
+        "]\n",
+        "\n",
+        "model_export_name = get_job_name_with_datetime(EXPORT_JOB_PREFIX + \"_\" + OBJECTIVE)\n",
+        "model_export_custom_job = aiplatform.CustomJob(\n",
+        "    display_name=model_export_name,\n",
+        "    project=PROJECT_ID,\n",
+        "    worker_pool_specs=worker_pool_specs,\n",
+        "    staging_bucket=STAGING_BUCKET,\n",
+        ")\n",
+        "\n",
+        "\n",
+        "model_export_custom_job.run()\n",
+        "\n",
+        "print(\"best model is saved to: \", os.path.join(model_dir, \"best_model\"))"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "g0BGaofgsMsy"
+      },
+      "source": [
+        "## Test trained models\n",
+        "This section shows how to test with trained models.\n",
+        "1. Upload and deploy models to model registry\n",
+        "2. Run predictions"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "NYuQowyZEtxK"
+      },
+      "outputs": [],
+      "source": [
+        "# @title Upload and deploy models\n",
+        "# model_dir is from the section above.\n",
+        "trained_model_dir = os.path.join(model_dir, \"best_model/saved_model\")\n",
+        "\n",
+        "upload_job_name = get_job_name_with_datetime(UPLOAD_JOB_PREFIX + \"_\" + OBJECTIVE)\n",
+        "\n",
+        "model = aiplatform.Model.upload(\n",
+        "    display_name=upload_job_name,\n",
+        "    artifact_uri=trained_model_dir,\n",
+        "    serving_container_image_uri=PREDICTION_CONTAINER_URI,\n",
+        "    serving_container_args=SERVING_CONTAINER_ARGS,\n",
+        ")\n",
+        "\n",
+        "model.wait()\n",
+        "\n",
+        "print(\"The uploaded model name is: \", upload_job_name)\n",
+        "\n",
+        "deploy_model_name = get_job_name_with_datetime(DEPLOY_JOB_PREFIX + \"_\" + OBJECTIVE)\n",
+        "print(\"The deployed job name is: \", deploy_model_name)\n",
+        "\n",
+        "endpoint = model.deploy(\n",
+        "    deployed_model_display_name=deploy_model_name,\n",
+        "    machine_type=PREDICTION_MACHINE_TYPE,\n",
+        "    traffic_split={\"0\": 100},\n",
+        "    accelerator_type=PREDICTION_ACCELERATOR_TYPE,\n",
+        "    accelerator_count=1,\n",
+        "    min_replica_count=1,\n",
+        "    max_replica_count=1,\n",
+        ")\n",
+        "\n",
+        "endpoint_id = endpoint.name\n",
+        "print(\"endpoint id is: \", endpoint_id)"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "vbIW9me1F2RY"
+      },
+      "outputs": [],
+      "source": [
+        "# @title Run predictions\n",
+        "\n",
+        "# endpoint_id was generated in the section above (`Upload and deploy models`).\n",
+        "endpoint_id = endpoint.name\n",
+        "\n",
+        "test_filepath = \"\"  # @param {type:\"string\"}\n",
+        "# If the input image is too large, we will resize it for prediction.\n",
+        "instances = get_prediction_instances(test_filepath, new_width=1000)\n",
+        "\n",
+        "# The label map file was generated from the section above (`Convert input data for training`).\n",
+        "label_map = get_label_map(label_map_path)[\"label_map\"]\n",
+        "\n",
+        "predictions, _ = predict_custom_trained_model(\n",
+        "    project=PROJECT_ID, location=REGION, endpoint_id=endpoint_id, instances=instances\n",
+        ")\n",
+        "\n",
+        "probs = dict(predictions[0])[\"probs\"]\n",
+        "max_prob = max(probs)\n",
+        "max_index = probs.index(max_prob)\n",
+        "print(\"The test image: \", test_filepath)\n",
+        "print(\"max_prob: \", max_prob, \", for label: \", label_map[max_index])\n",
+        "img = load_img(test_filepath)\n",
+        "display_image(img)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "kkH2nrpdp4sp"
+      },
+      "source": [
+        "## Clean up"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "Ax6vQVZhp9pR"
+      },
+      "outputs": [],
+      "source": [
+        "# Undeploy model and delete endpoint.\n",
+        "endpoint.delete(force=True)\n",
+        "# Delete models.\n",
+        "model.delete()\n",
+        "# Delete custom and hpt jobs.\n",
+        "if data_converter_custom_job.list(filter=f'display_name=\"{data_converter_job_name}\"'):\n",
+        "    data_converter_custom_job.delete()\n",
+        "if train_hpt_job.list(filter=f'display_name=\"{train_job_name}\"'):\n",
+        "    train_hpt_job.delete()\n",
+        "if model_export_custom_job.list(filter=f'display_name=\"{model_export_name}\"'):\n",
+        "    model_export_custom_job.delete()"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "name": "model_garden_tfvision_image_classification.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/community/model_garden/model_garden_tfvision_image_object_detection.ipynb

@@ -0,0 +1,975 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "ur8xi4C7S06n"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2023 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "TirJ-SGQseby"
+      },
+      "source": [
+        "# Vertex AI Model Garden TFVision With Image Object Detection\n",
+        "\n",
+        "<table align=\"left\">\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_tfvision_image_object_detection.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_tfvision_image_object_detection.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>                                                                                               <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/notebooks/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/main/notebooks/community/model_garden/model_garden_tfvision_image_object_detection.ipynb\">\n",
+        "      <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "dwGLvtIeECLK"
+      },
+      "source": [
+        "**_NOTE_**: This notebook has been tested in the following environment:\n",
+        "\n",
+        "* Python version = 3.9"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "tvgnzT1CKxrO"
+      },
+      "source": [
+        "## Overview\n",
+        "\n",
+        "This notebook demonstrates how to use [TFVision](https://github.com/tensorflow/models/blob/master/official/vision/MODEL_GARDEN.md) in Vertex AI Model Garden.\n",
+        "\n",
+        "### Objective\n",
+        "\n",
+        "* Train new models\n",
+        "  * Convert input data to training formats\n",
+        "  * Create [hyperparameter tuning jobs](https://cloud.google.com/vertex-ai/docs/training/hyperparameter-tuning-overview) to train new models\n",
+        "  * Find and export best models\n",
+        "\n",
+        "* Test trained models\n",
+        "  * Upload models to model registry\n",
+        "  * Deploy uploaded models\n",
+        "  * Run predictions\n",
+        "\n",
+        "* Cleanup resources\n",
+        "\n",
+        "### Costs\n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI\n",
+        "* Cloud Storage\n",
+        "\n",
+        "Learn about [Vertex AI\n",
+        "pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage\n",
+        "pricing](https://cloud.google.com/storage/pricing), and use the [Pricing\n",
+        "Calculator](https://cloud.google.com/products/calculator/)\n",
+        "to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "KEukV6uRk_S3"
+      },
+      "source": [
+        "## Before you begin"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "z__i0w0lCAsW"
+      },
+      "source": [
+        "### Colab Only\n",
+        "Run the following commands for colab and skip this section if you use workbench."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "Jvqs-ehKlaYh"
+      },
+      "outputs": [],
+      "source": [
+        "if \"google.colab\" in str(get_ipython()):\n",
+        "    ! pip3 install --upgrade google-cloud-aiplatform\n",
+        "\n",
+        "    # Automatically restart kernel after installs\n",
+        "    import IPython\n",
+        "\n",
+        "    app = IPython.Application.instance()\n",
+        "    app.kernel.do_shutdown(True)\n",
+        "\n",
+        "    from google.colab import auth as google_auth\n",
+        "\n",
+        "    google_auth.authenticate_user()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "BF1j6f9HApxa"
+      },
+      "source": [
+        "### Set up your Google Cloud project\n",
+        "\n",
+        "**The following steps are required, regardless of your notebook environment.**\n",
+        "\n",
+        "1. [Select or create a Google Cloud project](https://console.cloud.google.com/cloud-resource-manager). When you first create an account, you get a $300 free credit towards your compute/storage costs.\n",
+        "\n",
+        "1. [Make sure that billing is enabled for your project](https://cloud.google.com/billing/docs/how-to/modify-project).\n",
+        "\n",
+        "1. [Enable the Vertex AI API and Compute Engine API](https://console.cloud.google.com/flows/enableapi?apiid=aiplatform.googleapis.com,compute_component).\n",
+        "1. If you are running this notebook locally, you will need to install the [Cloud SDK](https://cloud.google.com/sdk).\n",
+        "\n",
+        "1. Enter your project ID in the cell below. Then run the cell to make sure the\n",
+        "Cloud SDK uses the right project for all the commands in this notebook.\n",
+        "\n",
+        "**Note**: Jupyter runs lines prefixed with `!` as shell commands, and it interpolates Python variables prefixed with `$` into these commands."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "9wExiMUxFk91"
+      },
+      "outputs": [],
+      "source": [
+        "import os\n",
+        "\n",
+        "from google.cloud import aiplatform\n",
+        "\n",
+        "# The project and bucket are for experiments below.\n",
+        "PROJECT_ID = \"\"  # @param {type:\"string\"}\n",
+        "BUCKET_URI = \"\"  # @param {type:\"string\"}\n",
+        "REGION = \"us-central1\"\n",
+        "\n",
+        "! gcloud config set project $PROJECT_ID\n",
+        "\n",
+        "STAGING_BUCKET = os.path.join(BUCKET_URI, \"temporal\")\n",
+        "\n",
+        "aiplatform.init(project=PROJECT_ID, location=REGION, staging_bucket=STAGING_BUCKET)\n",
+        "\n",
+        "# Download config files.\n",
+        "CONFIG_DIR = os.path.join(BUCKET_URI, \"config\")\n",
+        "! wget https://raw.githubusercontent.com/tensorflow/models/master/official/vision/configs/experiments/retinanet/coco_spinenet49_gpu_multiworker_mirrored.yaml\n",
+        "! gsutil cp coco_spinenet49_gpu_multiworker_mirrored.yaml $CONFIG_DIR\n",
+        "\n",
+        "! wget https://raw.githubusercontent.com/tensorflow/models/master/official/vision/configs/experiments/retinanet/coco_spinenet96_gpu_multiworker_mirrored.yaml\n",
+        "! gsutil cp coco_spinenet96_gpu_multiworker_mirrored.yaml $CONFIG_DIR\n",
+        "\n",
+        "! wget https://raw.githubusercontent.com/tensorflow/models/master/official/vision/configs/experiments/retinanet/coco_spinenet143_gpu_multiworker_mirrored.yaml\n",
+        "! gsutil cp coco_spinenet143_gpu_multiworker_mirrored.yaml $CONFIG_DIR\n",
+        "\n",
+        "! wget https://raw.githubusercontent.com/tensorflow/models/master/official/projects/yolo/configs/experiments/yolov4/detection/scaled_yolov4_1280_gpu.yaml\n",
+        "! gsutil cp scaled_yolov4_1280_gpu.yaml $CONFIG_DIR"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "n6IFz75WGCam"
+      },
+      "source": [
+        "### Define constants"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "riG_qUokg0XZ"
+      },
+      "outputs": [],
+      "source": [
+        "OBJECTIVE = \"iod\"\n",
+        "\n",
+        "# Data converter constants.\n",
+        "DATA_CONVERTER_JOB_PREFIX = \"data_converter\"\n",
+        "DATA_CONVERTER_CONTAINER = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/data-converter:latest\"\n",
+        "DATA_CONVERTER_MACHINE_TYPE = \"n1-highmem-8\"\n",
+        "\n",
+        "\n",
+        "# Training constants.\n",
+        "TRAINING_JOB_PREFIX = \"train\"\n",
+        "TRAIN_CONTAINER_URI = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/tfvision-oss:latest\"\n",
+        "TRAIN_MACHINE_TYPE = \"n1-highmem-16\"\n",
+        "TRAIN_ACCELERATOR_TYPE = \"NVIDIA_TESLA_V100\"\n",
+        "TRAIN_NUM_GPU = 2\n",
+        "TRAIN_SPINENET49_CONFIG = os.path.join(\n",
+        "    CONFIG_DIR, \"coco_spinenet49_gpu_multiworker_mirrored.yaml\"\n",
+        ")\n",
+        "TRAIN_SPINENET96_CONFIG = os.path.join(\n",
+        "    CONFIG_DIR, \"coco_spinenet96_gpu_multiworker_mirrored.yaml\"\n",
+        ")\n",
+        "TRAIN_SPINENET143_CONFIG = os.path.join(\n",
+        "    CONFIG_DIR, \"coco_spinenet143_gpu_multiworker_mirrored.yaml\"\n",
+        ")\n",
+        "TRAIN_YOLOV4_CONFIG = os.path.join(CONFIG_DIR, \"scaled_yolov4_1280_gpu.yaml\")\n",
+        "\n",
+        "# Evaluation constants.\n",
+        "EVALUATION_METRIC = \"AP50\"\n",
+        "\n",
+        "# Export constants.\n",
+        "EXPORT_JOB_PREFIX = \"export\"\n",
+        "EXPORT_CONTAINER_URI = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/tfvision-serving:latest\"\n",
+        "EXPORT_MACHINE_TYPE = \"n1-highmem-8\"\n",
+        "\n",
+        "# Prediction constants.\n",
+        "# You can deploy models with\n",
+        "#   pre-build-dockers: https://cloud.google.com/vertex-ai/docs/predictions/pre-built-containers.\n",
+        "#   and optimized tensorflow runtime dockers: https://cloud.google.com/vertex-ai/docs/predictions/optimized-tensorflow-runtime.\n",
+        "# The example in this notebook uses optimized tensorflow runtime dockers.\n",
+        "# You can adjust accelerator types and machine types to get faster predictions.\n",
+        "PREDICTION_CONTAINER_URI = (\n",
+        "    \"us-docker.pkg.dev/vertex-ai-restricted/prediction/tf_opt-gpu.2-11:latest\"\n",
+        ")\n",
+        "SERVING_CONTAINER_ARGS = [\"--allow_precompilation\", \"--allow_compression\"]\n",
+        "PREDICTION_ACCELERATOR_TYPE = \"NVIDIA_TESLA_T4\"\n",
+        "PREDICTION_MACHINE_TYPE = \"n1-standard-4\"\n",
+        "UPLOAD_JOB_PREFIX = \"upload\"\n",
+        "DEPLOY_JOB_PREFIX = \"deploy\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "ZZFPe_GezXg8"
+      },
+      "source": [
+        "### Define common libraries"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "XcYUGwr-AJGY"
+      },
+      "outputs": [],
+      "source": [
+        "import base64\n",
+        "import json\n",
+        "import os\n",
+        "from datetime import datetime\n",
+        "from io import BytesIO\n",
+        "from typing import Dict, List, Union\n",
+        "\n",
+        "import matplotlib.pyplot as plt\n",
+        "import numpy as np\n",
+        "import tensorflow as tf\n",
+        "import yaml\n",
+        "from google.cloud import aiplatform\n",
+        "from google.protobuf import json_format\n",
+        "from google.protobuf.struct_pb2 import Value\n",
+        "from PIL import Image, ImageColor, ImageDraw, ImageFont\n",
+        "\n",
+        "\n",
+        "def get_job_name_with_datetime(prefix: str):\n",
+        "    return prefix + datetime.now().strftime(\"_%Y%m%d_%H%M%S\")\n",
+        "\n",
+        "\n",
+        "def predict_custom_trained_model(\n",
+        "    project: str,\n",
+        "    endpoint_id: str,\n",
+        "    instances: Union[Dict, List[Dict]],\n",
+        "    location: str = \"us-central1\",\n",
+        "    api_endpoint: str = \"us-central1-aiplatform.googleapis.com\",\n",
+        "):\n",
+        "    # The AI Platform services require regional API endpoints.\n",
+        "    client_options = {\"api_endpoint\": api_endpoint}\n",
+        "    # Initialize client that will be used to create and send requests.\n",
+        "    # This client only needs to be created once, and can be reused for multiple requests.\n",
+        "    client = aiplatform.gapic.PredictionServiceClient(client_options=client_options)\n",
+        "    parameters_dict = {}\n",
+        "    parameters = json_format.ParseDict(parameters_dict, Value())\n",
+        "    endpoint = client.endpoint_path(\n",
+        "        project=project, location=location, endpoint=endpoint_id\n",
+        "    )\n",
+        "    response = client.predict(\n",
+        "        endpoint=endpoint, instances=instances, parameters=parameters\n",
+        "    )\n",
+        "    return response.predictions, response.deployed_model_id\n",
+        "\n",
+        "\n",
+        "def load_img(path):\n",
+        "    img = tf.io.read_file(path)\n",
+        "    img = tf.image.decode_jpeg(img, channels=3)\n",
+        "    return Image.fromarray(np.uint8(img)).convert(\"RGB\")\n",
+        "\n",
+        "\n",
+        "def display_image(image):\n",
+        "    _ = plt.figure(figsize=(20, 15))\n",
+        "    plt.grid(False)\n",
+        "    plt.imshow(image)\n",
+        "\n",
+        "\n",
+        "def get_prediction_instances(test_filepath, new_width=-1):\n",
+        "    if new_width <= 0:\n",
+        "        with tf.io.gfile.GFile(test_filepath, \"rb\") as input_file:\n",
+        "            encoded_string = base64.b64encode(input_file.read()).decode(\"utf-8\")\n",
+        "    else:\n",
+        "        img = load_img(test_filepath)\n",
+        "        width, height = img.size\n",
+        "        print(\"original input image size: \", width, \" , \", height)\n",
+        "        new_height = int(height * new_width / width)\n",
+        "        new_img = img.resize((new_width, new_height))\n",
+        "        print(\"resized input image size: \", new_width, \" , \", new_height)\n",
+        "        buffered = BytesIO()\n",
+        "        new_img.save(buffered, format=\"JPEG\")\n",
+        "        encoded_string = base64.b64encode(buffered.getvalue()).decode(\"utf-8\")\n",
+        "\n",
+        "    instances = [\n",
+        "        {\n",
+        "            \"encoded_image\": {\"b64\": encoded_string},\n",
+        "        }\n",
+        "    ]\n",
+        "    return instances\n",
+        "\n",
+        "\n",
+        "def get_label_map(label_map_yaml_filepath):\n",
+        "    with tf.io.gfile.GFile(label_map_yaml_filepath, \"rb\") as input_file:\n",
+        "        label_map = yaml.safe_load(input_file.read())\n",
+        "    return label_map\n",
+        "\n",
+        "\n",
+        "def get_best_trial(model_dir, max_trial_count, evaluation_metric):\n",
+        "    best_trial_dir = \"\"\n",
+        "    best_trial_evaluation_results = {}\n",
+        "    best_performance = -1\n",
+        "\n",
+        "    for i in range(max_trial_count):\n",
+        "        current_trial = i + 1\n",
+        "        current_trial_dir = os.path.join(model_dir, \"trial_\" + str(current_trial))\n",
+        "        current_trial_best_ckpt_dir = os.path.join(current_trial_dir, \"best_ckpt\")\n",
+        "        current_trial_best_ckpt_evaluation_filepath = os.path.join(\n",
+        "            current_trial_best_ckpt_dir, \"info.json\"\n",
+        "        )\n",
+        "        with tf.io.gfile.GFile(current_trial_best_ckpt_evaluation_filepath, \"rb\") as f:\n",
+        "            eval_metric_results = json.load(f)\n",
+        "            current_performance = eval_metric_results[evaluation_metric]\n",
+        "            if current_performance > best_performance:\n",
+        "                best_performance = current_performance\n",
+        "                best_trial_dir = current_trial_dir\n",
+        "                best_trial_evaluation_results = eval_metric_results\n",
+        "    return best_trial_dir, best_trial_evaluation_results\n",
+        "\n",
+        "\n",
+        "def draw_bounding_box_on_image(\n",
+        "    image, ymin, xmin, ymax, xmax, color, font, thickness=4, display_str_list=()\n",
+        "):\n",
+        "    \"\"\"Adds a bounding box to an image.\"\"\"\n",
+        "    draw = ImageDraw.Draw(image)\n",
+        "    im_width, im_height = image.size\n",
+        "    (left, right, top, bottom) = (\n",
+        "        xmin * im_width,\n",
+        "        xmax * im_width,\n",
+        "        ymin * im_height,\n",
+        "        ymax * im_height,\n",
+        "    )\n",
+        "    draw.line(\n",
+        "        [(left, top), (left, bottom), (right, bottom), (right, top), (left, top)],\n",
+        "        width=thickness,\n",
+        "        fill=color,\n",
+        "    )\n",
+        "\n",
+        "    # If the total height of the display strings added to the top of the bounding\n",
+        "    # box exceeds the top of the image, stack the strings below the bounding box\n",
+        "    # instead of above.\n",
+        "    display_str_heights = [font.getsize(ds)[1] for ds in display_str_list]\n",
+        "    # Each display_str has a top and bottom margin of 0.05x.\n",
+        "    total_display_str_height = (1 + 2 * 0.05) * sum(display_str_heights)\n",
+        "\n",
+        "    if top > total_display_str_height:\n",
+        "        text_bottom = top\n",
+        "    else:\n",
+        "        text_bottom = top + total_display_str_height\n",
+        "    # Reverse list and print from bottom to top.\n",
+        "    for display_str in display_str_list[::-1]:\n",
+        "        text_width, text_height = font.getsize(display_str)\n",
+        "        margin = np.ceil(0.05 * text_height)\n",
+        "        draw.rectangle(\n",
+        "            [\n",
+        "                (left, text_bottom - text_height - 2 * margin),\n",
+        "                (left + text_width, text_bottom),\n",
+        "            ],\n",
+        "            fill=color,\n",
+        "        )\n",
+        "        draw.text(\n",
+        "            (left + margin, text_bottom - text_height - margin),\n",
+        "            display_str,\n",
+        "            fill=\"black\",\n",
+        "            font=font,\n",
+        "        )\n",
+        "        text_bottom -= text_height - 2 * margin\n",
+        "\n",
+        "\n",
+        "def draw_boxes(image, boxes, class_names, scores, max_boxes=40, min_score=0.05):\n",
+        "    \"\"\"Overlay labeled boxes on an image with formatted scores and label names.\"\"\"\n",
+        "    colors = list(ImageColor.colormap.values())\n",
+        "    try:\n",
+        "        font = ImageFont.truetype(\n",
+        "            \"/usr/share/fonts/truetype/liberation/LiberationSansNarrow-Regular.ttf\", 25\n",
+        "        )\n",
+        "    except IOError:\n",
+        "        print(\"Font not found, using default font.\")\n",
+        "        font = ImageFont.load_default()\n",
+        "\n",
+        "    for i in range(min(len(boxes), max_boxes)):\n",
+        "        if scores[i] >= min_score:\n",
+        "            ymin, xmin, ymax, xmax = boxes[i]\n",
+        "            display_str = \"{}: {}%\".format(class_names[i], int(100 * scores[i]))\n",
+        "            color = colors[hash(class_names[i]) % len(colors)]\n",
+        "            draw_bounding_box_on_image(\n",
+        "                image,\n",
+        "                ymin,\n",
+        "                xmin,\n",
+        "                ymax,\n",
+        "                xmax,\n",
+        "                color,\n",
+        "                font,\n",
+        "                display_str_list=[display_str],\n",
+        "            )\n",
+        "    return image"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "RB_xY9ipr7ZU"
+      },
+      "source": [
+        "## Train new models\n",
+        "This section shows how to train new models.\n",
+        "1. Convert input data to training formats\n",
+        "2. Create hyperparameter tuning jobs to train new models\n",
+        "3. Find and export best models\n",
+        "\n",
+        "If you already trained models, please go to the section `Test trained models`."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "zgPO1eR3CYjk"
+      },
+      "source": [
+        "### Prepare input data for training\n",
+        "\n",
+        "Prepare data in the format as described [here](https://cloud.google.com/vertex-ai/docs/image-data/classification/prepare-data), and then convert them to the training formats as below:\n",
+        "\n",
+        "* `input_file_path`: The input file path for preparing data.\n",
+        "* `input_file_type`: The input file type, such as csv or jsonl.\n",
+        "* `split_ratio`: The proportion of data to split into train/validation/test.\n",
+        "* `num_shard`: The number of shards for train/validation/test.\n",
+        "* `output_dir`: The output directory, which will container prepared train/test/validation data."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "IndQ_m6ddUEM"
+      },
+      "outputs": [],
+      "source": [
+        "# This job will convert input data as training format, with given split ratios\n",
+        "# and number of shards on train/test/validation.\n",
+        "data_converter_job_name = get_job_name_with_datetime(\n",
+        "    DATA_CONVERTER_JOB_PREFIX + \"_\" + OBJECTIVE\n",
+        ")\n",
+        "\n",
+        "input_file_path = \"\"  # @param {type:\"string\"}\n",
+        "input_file_type = \"\"  # @param {type:\"string\"}\n",
+        "split_ratio = \"0.8,0.1,0.1\"\n",
+        "num_shard = \"10,10,10\"\n",
+        "data_converter_output_dir = os.path.join(BUCKET_URI, data_converter_job_name)\n",
+        "\n",
+        "\n",
+        "worker_pool_specs = [\n",
+        "    {\n",
+        "        \"machine_spec\": {\n",
+        "            \"machine_type\": DATA_CONVERTER_MACHINE_TYPE,\n",
+        "        },\n",
+        "        \"replica_count\": 1,\n",
+        "        \"container_spec\": {\n",
+        "            \"image_uri\": DATA_CONVERTER_CONTAINER,\n",
+        "            \"command\": [],\n",
+        "            \"args\": [\n",
+        "                \"--input_file_path=%s\" % input_file_path,\n",
+        "                \"--input_file_type=%s\" % input_file_type,\n",
+        "                \"--objective=%s\" % OBJECTIVE,\n",
+        "                \"--num_shard=%s\" % num_shard,\n",
+        "                \"--split_ratio=%s\" % split_ratio,\n",
+        "                \"--output_dir=%s\" % data_converter_output_dir,\n",
+        "            ],\n",
+        "        },\n",
+        "    }\n",
+        "]\n",
+        "\n",
+        "data_converter_custom_job = aiplatform.CustomJob(\n",
+        "    display_name=data_converter_job_name,\n",
+        "    project=PROJECT_ID,\n",
+        "    worker_pool_specs=worker_pool_specs,\n",
+        "    staging_bucket=STAGING_BUCKET,\n",
+        ")\n",
+        "\n",
+        "data_converter_custom_job.run()\n",
+        "\n",
+        "input_train_data_path = os.path.join(data_converter_output_dir, \"train.tfrecord*\")\n",
+        "input_validation_data_path = os.path.join(data_converter_output_dir, \"val.tfrecord*\")\n",
+        "label_map_path = os.path.join(data_converter_output_dir, \"label_map.yaml\")\n",
+        "print(\"input_train_data_path for training: \", input_train_data_path)\n",
+        "print(\"input_validation_data_path for training: \", input_validation_data_path)\n",
+        "print(\"label_map_path for prediction: \", label_map_path)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "SA8DVTn7j69v"
+      },
+      "source": [
+        "### Create a Vertex AI custom job with hyperparameter tuning\n",
+        "\n",
+        "You use the Vertex AI SDK to create and run the hyperparameter tuning job with Vertex AI Model Garden Training Dockers."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "aaff6f5be7f6"
+      },
+      "source": [
+        "#### Define the following specifications\n",
+        "\n",
+        "* `worker_pool_specs`: Dictionary specifying the machine type and Docker image. This example defines a single node cluster with one `n1-standard-4` machine with two `NVIDIA_TESLA_T4` GPUs.\n",
+        "* `parameter_spec`: Dictionary specifying the parameters to optimize. The dictionary key is the string assigned to the command line argument for each hyperparameter in your training application code, and the dictionary value is the parameter specification. The parameter specification includes the type, min/max values, and scale for the hyperparameter.\n",
+        "* `metric_spec`: Dictionary specifying the metric to optimize. The dictionary key is the `hyperparameter_metric_tag` that you set in your training application code, and the value is the optimization goal."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "um_XKbmpTaHx"
+      },
+      "outputs": [],
+      "source": [
+        "from google.cloud.aiplatform import hyperparameter_tuning as hpt\n",
+        "\n",
+        "label_map = get_label_map(label_map_path)\n",
+        "num_classes = len(label_map[\"label_map\"]) + 1\n",
+        "\n",
+        "# Input train and validation datasets can be found from the section above\n",
+        "# `Convert input data for training`.\n",
+        "# Set prepared datasets if exists.\n",
+        "# input_train_data_path = ''\n",
+        "# input_validation_data_path = ''\n",
+        "\n",
+        "# Refer to https://github.com/tensorflow/models/blob/master/official/vision/MODEL_GARDEN.md\n",
+        "# for more model details.\n",
+        "experiment = \"retinanet_spinenet96\"  # @param ['retinanet_spinenet49', \"retinanet_spinenet96\", 'retinanet_spinenet143', 'scaled_yolo_v4']\n",
+        "\n",
+        "train_job_name = get_job_name_with_datetime(TRAINING_JOB_PREFIX + \"_\" + OBJECTIVE)\n",
+        "model_dir = os.path.join(BUCKET_URI, train_job_name)\n",
+        "\n",
+        "# The arguments here are mainly for test purposes. Please update them\n",
+        "# to get better performances.\n",
+        "common_args = {\n",
+        "    \"input_train_data_path\": input_train_data_path,\n",
+        "    \"input_validation_data_path\": input_validation_data_path,\n",
+        "    \"objective\": OBJECTIVE,\n",
+        "    \"model_dir\": model_dir,\n",
+        "    \"num_classes\": num_classes,\n",
+        "    \"global_batch_size\": 4,\n",
+        "    \"prefetch_buffer_size\": 12,\n",
+        "    \"train_steps\": 2000,\n",
+        "    \"input_size\": \"1024,1024\",\n",
+        "}\n",
+        "\n",
+        "experiment_container_args_dict = {\n",
+        "    # retinanet_spinenet49 experiment args.\n",
+        "    \"retinanet_spinenet49\": dict(\n",
+        "        common_args,\n",
+        "        **{\n",
+        "            \"experiment\": \"retinanet_spinenet_coco\",\n",
+        "            \"config_file\": TRAIN_SPINENET49_CONFIG,\n",
+        "            \"anchor_size\": 4,\n",
+        "        },\n",
+        "    ),\n",
+        "    # retinanet_spinenet96 experiment args.\n",
+        "    \"retinanet_spinenet96\": dict(\n",
+        "        common_args,\n",
+        "        **{\n",
+        "            \"experiment\": \"retinanet_spinenet_coco\",\n",
+        "            \"config_file\": TRAIN_SPINENET96_CONFIG,\n",
+        "            \"anchor_size\": 4,\n",
+        "        },\n",
+        "    ),\n",
+        "    # retinanet_spinenet143 experiment args.\n",
+        "    \"retinanet_spinenet143\": dict(\n",
+        "        common_args,\n",
+        "        **{\n",
+        "            \"experiment\": \"retinanet_spinenet_coco\",\n",
+        "            \"config_file\": TRAIN_SPINENET96_CONFIG,\n",
+        "            \"anchor_size\": 4,\n",
+        "        },\n",
+        "    ),\n",
+        "    # scaled_yolo_v4 experiment args.\n",
+        "    \"scaled_yolo_v4\": dict(\n",
+        "        common_args,\n",
+        "        **{\n",
+        "            \"experiment\": \"scaled_yolo\",\n",
+        "            \"config_file\": TRAIN_YOLOV4_CONFIG,\n",
+        "        },\n",
+        "    ),\n",
+        "}\n",
+        "\n",
+        "params_override = \"runtime.num_gpus=%s\" % TRAIN_NUM_GPU\n",
+        "eval_params_override = \"runtime.num_gpus=1,runtime.distribution_strategy=mirrored\"\n",
+        "worker_pool_specs = [\n",
+        "    {\n",
+        "        \"machine_spec\": {\n",
+        "            \"machine_type\": TRAIN_MACHINE_TYPE,\n",
+        "            \"accelerator_type\": TRAIN_ACCELERATOR_TYPE,\n",
+        "            \"accelerator_count\": TRAIN_NUM_GPU,\n",
+        "        },\n",
+        "        \"replica_count\": 1,\n",
+        "        \"container_spec\": {\n",
+        "            \"image_uri\": TRAIN_CONTAINER_URI,\n",
+        "            \"args\": [\n",
+        "                \"--mode=train\",\n",
+        "                \"--params_override=%s\" % params_override,\n",
+        "            ]\n",
+        "            + [\n",
+        "                \"--{}={}\".format(k, v)\n",
+        "                for k, v in experiment_container_args_dict[experiment].items()\n",
+        "            ],\n",
+        "        },\n",
+        "    },\n",
+        "    {},\n",
+        "    {},\n",
+        "    {\n",
+        "        \"machine_spec\": {\n",
+        "            \"machine_type\": \"n1-highmem-4\",\n",
+        "            \"accelerator_type\": TRAIN_ACCELERATOR_TYPE,\n",
+        "            \"accelerator_count\": 1,\n",
+        "        },\n",
+        "        \"replica_count\": 1,\n",
+        "        \"container_spec\": {\n",
+        "            \"image_uri\": TRAIN_CONTAINER_URI,\n",
+        "            \"args\": [\n",
+        "                \"--mode=continuous_eval\",\n",
+        "                \"--params_override=%s\" % eval_params_override,\n",
+        "            ]\n",
+        "            + [\n",
+        "                \"--{}={}\".format(k, v)\n",
+        "                for k, v in experiment_container_args_dict[experiment].items()\n",
+        "            ],\n",
+        "        },\n",
+        "    },\n",
+        "]\n",
+        "\n",
+        "metric_spec = {\"model_performance\": \"maximize\"}\n",
+        "\n",
+        "LEARNING_RATES = [0.001, 0.01]\n",
+        "# Models will be trained with each learning rate separately and max trial count is the number of learning rates.\n",
+        "MAX_TRIAL_COUNT = len(LEARNING_RATES)\n",
+        "parameter_spec = {\n",
+        "    \"learning_rate\": hpt.DiscreteParameterSpec(values=LEARNING_RATES, scale=\"linear\"),\n",
+        "}\n",
+        "\n",
+        "print(worker_pool_specs, metric_spec, parameter_spec)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "HwcCjwlBTQIz"
+      },
+      "source": [
+        "#### Run hyperparameter tuning jobs\n",
+        "* `max_trial_count`: Sets an upper bound on the number of trials the service will run. The recommended practice is to start with a smaller number of trials and get a sense of how impactful your chosen hyperparameters are before scaling up.\n",
+        "\n",
+        "* `parallel_trial_count`:  If you use parallel trials, the service provisions multiple training processing clusters. The worker pool spec that you specify when creating the job is used for each individual training cluster.  Increasing the number of parallel trials reduces the amount of time the hyperparameter tuning job takes to run; however, it can reduce the effectiveness of the job overall. This is because the default tuning strategy uses results of previous trials to inform the assignment of values in subsequent trials.\n",
+        "\n",
+        "* `search_algorithm`: The available search algorithms are grid, random, or default (None). The default option applies Bayesian optimization to search the space of possible hyperparameter values and is the recommended algorithm.\n",
+        "\n",
+        "Click on the generated link in the output to see your run in the Cloud Console."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "aec22792ee84"
+      },
+      "outputs": [],
+      "source": [
+        "train_custom_job = aiplatform.CustomJob(\n",
+        "    display_name=train_job_name,\n",
+        "    project=PROJECT_ID,\n",
+        "    worker_pool_specs=worker_pool_specs,\n",
+        "    staging_bucket=STAGING_BUCKET,\n",
+        ")\n",
+        "\n",
+        "train_hpt_job = aiplatform.HyperparameterTuningJob(\n",
+        "    display_name=train_job_name,\n",
+        "    custom_job=train_custom_job,\n",
+        "    metric_spec=metric_spec,\n",
+        "    parameter_spec=parameter_spec,\n",
+        "    max_trial_count=MAX_TRIAL_COUNT,\n",
+        "    parallel_trial_count=1,\n",
+        "    project=PROJECT_ID,\n",
+        "    search_algorithm=None,\n",
+        ")\n",
+        "\n",
+        "train_hpt_job.run()\n",
+        "\n",
+        "print(\"experiment is: \", experiment)\n",
+        "print(\"model_dir is: \", model_dir)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "mV-Djz-frBni"
+      },
+      "source": [
+        "### Export best models as TF Saved Model format"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "09Rz1AYspK19"
+      },
+      "outputs": [],
+      "source": [
+        "# This job will export models from TF checkpoints to TF saved model format.\n",
+        "# model_dir is from the section above.\n",
+        "best_trial_dir, best_trial_evaluation_results = get_best_trial(\n",
+        "    model_dir, MAX_TRIAL_COUNT, EVALUATION_METRIC\n",
+        ")\n",
+        "print(\"best_trial_dir: \", best_trial_dir)\n",
+        "print(\"best_trial_evaluation_results: \", best_trial_evaluation_results)\n",
+        "\n",
+        "worker_pool_specs = [\n",
+        "    {\n",
+        "        \"machine_spec\": {\n",
+        "            \"machine_type\": EXPORT_MACHINE_TYPE,\n",
+        "        },\n",
+        "        \"replica_count\": 1,\n",
+        "        \"container_spec\": {\n",
+        "            \"image_uri\": EXPORT_CONTAINER_URI,\n",
+        "            \"command\": [],\n",
+        "            \"args\": [\n",
+        "                \"--objective=%s\" % OBJECTIVE,\n",
+        "                \"--input_image_size=1024,1024\",\n",
+        "                \"--experiment=%s\"\n",
+        "                % experiment_container_args_dict[experiment][\"experiment\"],\n",
+        "                \"--config_file=%s/params.yaml\" % best_trial_dir,\n",
+        "                \"--checkpoint_path=%s/best_ckpt\" % best_trial_dir,\n",
+        "                \"--export_dir=%s/best_model\" % model_dir,\n",
+        "            ],\n",
+        "        },\n",
+        "    }\n",
+        "]\n",
+        "\n",
+        "model_export_name = get_job_name_with_datetime(EXPORT_JOB_PREFIX + \"_\" + OBJECTIVE)\n",
+        "model_export_custom_job = aiplatform.CustomJob(\n",
+        "    display_name=model_export_name,\n",
+        "    project=PROJECT_ID,\n",
+        "    worker_pool_specs=worker_pool_specs,\n",
+        "    staging_bucket=STAGING_BUCKET,\n",
+        ")\n",
+        "\n",
+        "\n",
+        "model_export_custom_job.run()\n",
+        "\n",
+        "print(\"best model is saved to: \", os.path.join(model_dir, \"best_model\"))"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "g0BGaofgsMsy"
+      },
+      "source": [
+        "## Test trained models\n",
+        "This section will show how to test with trained models.\n",
+        "1. Upload and deploy models\n",
+        "2. Run predictions"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "NYuQowyZEtxK"
+      },
+      "outputs": [],
+      "source": [
+        "# @title Upload and deploy models\n",
+        "# model_dir is from the section above.\n",
+        "trained_model_dir = os.path.join(model_dir, \"best_model/saved_model\")\n",
+        "\n",
+        "upload_job_name = get_job_name_with_datetime(UPLOAD_JOB_PREFIX + \"_\" + OBJECTIVE)\n",
+        "\n",
+        "model = aiplatform.Model.upload(\n",
+        "    display_name=upload_job_name,\n",
+        "    artifact_uri=trained_model_dir,\n",
+        "    serving_container_image_uri=PREDICTION_CONTAINER_URI,\n",
+        "    serving_container_args=SERVING_CONTAINER_ARGS,\n",
+        ")\n",
+        "\n",
+        "model.wait()\n",
+        "\n",
+        "print(\"The uploaded model name is: \", upload_job_name)\n",
+        "\n",
+        "deploy_model_name = get_job_name_with_datetime(DEPLOY_JOB_PREFIX + \"_\" + OBJECTIVE)\n",
+        "print(\"The deployed job name is: \", deploy_model_name)\n",
+        "\n",
+        "endpoint = model.deploy(\n",
+        "    deployed_model_display_name=deploy_model_name,\n",
+        "    machine_type=PREDICTION_MACHINE_TYPE,\n",
+        "    traffic_split={\"0\": 100},\n",
+        "    accelerator_type=PREDICTION_ACCELERATOR_TYPE,\n",
+        "    accelerator_count=1,\n",
+        "    min_replica_count=1,\n",
+        "    max_replica_count=1,\n",
+        ")\n",
+        "\n",
+        "endpoint_id = endpoint.name\n",
+        "print(\"endpoint id is: \", endpoint_id)"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "vbIW9me1F2RY"
+      },
+      "outputs": [],
+      "source": [
+        "# @title Run predictions\n",
+        "\n",
+        "# endpoint_id was generated in the section above (`Upload and deploy models`).\n",
+        "endpoint_id = endpoint.name\n",
+        "\n",
+        "# The test image file path.\n",
+        "test_filepath = \"\"  # @param {type:\"string\"}\n",
+        "score_threshold = 0.2  # @param {type:\"number\"}\n",
+        "# If the input image is too large, we will resize it for prediction.\n",
+        "instances = get_prediction_instances(test_filepath, new_width=1000)\n",
+        "\n",
+        "# The label map file was generated from the section above (`Convert input data for training`).\n",
+        "label_map = get_label_map(label_map_path)[\"label_map\"]\n",
+        "\n",
+        "predictions, _ = predict_custom_trained_model(\n",
+        "    project=PROJECT_ID, location=REGION, endpoint_id=endpoint_id, instances=instances\n",
+        ")\n",
+        "\n",
+        "img = load_img(test_filepath)\n",
+        "detection_boxes = predictions[0][\"detection_boxes\"]\n",
+        "detection_scores = predictions[0][\"detection_scores\"]\n",
+        "detection_classes_as_text = []\n",
+        "\n",
+        "for detection_class in predictions[0][\"detection_classes\"]:\n",
+        "    detection_classes_as_text.append(label_map[int(detection_class)])\n",
+        "\n",
+        "img = draw_boxes(\n",
+        "    img,\n",
+        "    detection_boxes,\n",
+        "    detection_classes_as_text,\n",
+        "    detection_scores,\n",
+        "    min_score=score_threshold,\n",
+        ")\n",
+        "display_image(img)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "kkH2nrpdp4sp"
+      },
+      "source": [
+        "## Clean up"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "Ax6vQVZhp9pR"
+      },
+      "outputs": [],
+      "source": [
+        "# Undeploy model and delete endpoint.\n",
+        "endpoint.delete(force=True)\n",
+        "# Delete models.\n",
+        "model.delete()\n",
+        "# Delete custom and hpt jobs.\n",
+        "if data_converter_custom_job.list(filter=f'display_name=\"{data_converter_job_name}\"'):\n",
+        "    data_converter_custom_job.delete()\n",
+        "if train_hpt_job.list(filter=f'display_name=\"{train_job_name}\"'):\n",
+        "    train_hpt_job.delete()\n",
+        "if model_export_custom_job.list(filter=f'display_name=\"{model_export_name}\"'):\n",
+        "    model_export_custom_job.delete()"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "name": "model_garden_tfvision_image_object_detection.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/community/model_garden/model_garden_tfvision_image_segmentation.ipynb

@@ -0,0 +1,946 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "ur8xi4C7S06n"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2023 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "TirJ-SGQseby"
+      },
+      "source": [
+        "# Vertex AI Model Garden TFVision With Image Segmentation\n",
+        "\n",
+        "<table align=\"left\">\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_tfvision_image_segmentation.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_tfvision_image_segmentation.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>                                                                                               <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/notebooks/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/main/notebooks/community/model_garden/model_garden_tfvision_image_segmentation.ipynb\">\n",
+        "      <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "dwGLvtIeECLK"
+      },
+      "source": [
+        "**_NOTE_**: This notebook has been tested in the following environment:\n",
+        "\n",
+        "* Python version = 3.9"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "tvgnzT1CKxrO"
+      },
+      "source": [
+        "## Overview\n",
+        "\n",
+        "This notebook demonstrates how to use [TFVision](https://github.com/tensorflow/models/blob/master/official/vision/MODEL_GARDEN.md) in Vertex AI Model Garden.\n",
+        "\n",
+        "### Objective\n",
+        "\n",
+        "* Train new models\n",
+        "  * Convert input data to training formats\n",
+        "  * Create [hyperparameter tuning jobs](https://cloud.google.com/vertex-ai/docs/training/hyperparameter-tuning-overview) to train new models\n",
+        "  * Find and export best models\n",
+        "\n",
+        "* Test trained models\n",
+        "  * Upload models to model registry\n",
+        "  * Deploy uploaded models\n",
+        "  * Run predictions\n",
+        "\n",
+        "* Cleanup resources\n",
+        "\n",
+        "### Costs\n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI\n",
+        "* Cloud Storage\n",
+        "\n",
+        "Learn about [Vertex AI\n",
+        "pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage\n",
+        "pricing](https://cloud.google.com/storage/pricing), and use the [Pricing\n",
+        "Calculator](https://cloud.google.com/products/calculator/)\n",
+        "to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "KEukV6uRk_S3"
+      },
+      "source": [
+        "## Before you begin"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "z__i0w0lCAsW"
+      },
+      "source": [
+        "### Colab Only\n",
+        "\n",
+        "Run the following commands for colab and skip this section if you use workbench."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "Jvqs-ehKlaYh"
+      },
+      "outputs": [],
+      "source": [
+        "if \"google.colab\" in str(get_ipython()):\n",
+        "    ! pip3 install --upgrade google-cloud-aiplatform\n",
+        "\n",
+        "    # Automatically restart kernel after installs\n",
+        "    import IPython\n",
+        "\n",
+        "    app = IPython.Application.instance()\n",
+        "    app.kernel.do_shutdown(True)\n",
+        "\n",
+        "    from google.colab import auth as google_auth\n",
+        "\n",
+        "    google_auth.authenticate_user()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "BF1j6f9HApxa"
+      },
+      "source": [
+        "### Set up your Google Cloud project\n",
+        "\n",
+        "**The following steps are required, regardless of your notebook environment.**\n",
+        "\n",
+        "1. [Select or create a Google Cloud project](https://console.cloud.google.com/cloud-resource-manager). When you first create an account, you get a $300 free credit towards your compute/storage costs.\n",
+        "\n",
+        "1. [Make sure that billing is enabled for your project](https://cloud.google.com/billing/docs/how-to/modify-project).\n",
+        "\n",
+        "1. [Enable the Vertex AI API and Compute Engine API](https://console.cloud.google.com/flows/enableapi?apiid=aiplatform.googleapis.com,compute_component).\n",
+        "1. If you are running this notebook locally, you will need to install the [Cloud SDK](https://cloud.google.com/sdk).\n",
+        "\n",
+        "1. Enter your project ID in the cell below. Then run the cell to make sure the\n",
+        "Cloud SDK uses the right project for all the commands in this notebook.\n",
+        "\n",
+        "**Note**: Jupyter runs lines prefixed with `!` as shell commands, and it interpolates Python variables prefixed with `$` into these commands."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "9wExiMUxFk91"
+      },
+      "outputs": [],
+      "source": [
+        "import os\n",
+        "\n",
+        "from google.cloud import aiplatform\n",
+        "\n",
+        "# The project and bucket are for experiments below.\n",
+        "PROJECT_ID = \"\"  # @param {type:\"string\"}\n",
+        "BUCKET_URI = \"\"  # @param {type:\"string\"}\n",
+        "REGION = \"us-central1\"\n",
+        "\n",
+        "! gcloud config set project $PROJECT_ID\n",
+        "\n",
+        "STAGING_BUCKET = os.path.join(BUCKET_URI, \"temporal\")\n",
+        "\n",
+        "aiplatform.init(project=PROJECT_ID, location=REGION, staging_bucket=STAGING_BUCKET)\n",
+        "\n",
+        "# Download config files.\n",
+        "CONFIG_DIR = os.path.join(BUCKET_URI, \"config\")\n",
+        "! wget https://raw.githubusercontent.com/tensorflow/models/master/official/vision/configs/experiments/semantic_segmentation/deeplabv3plus_resnet101_cityscapes_gpu_multiworker_mirrored.yaml\n",
+        "! gsutil cp deeplabv3plus_resnet101_cityscapes_gpu_multiworker_mirrored.yaml $CONFIG_DIR"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "n6IFz75WGCam"
+      },
+      "source": [
+        "### Define constants"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "riG_qUokg0XZ"
+      },
+      "outputs": [],
+      "source": [
+        "OBJECTIVE = \"isg\"\n",
+        "\n",
+        "# Data converter constants.\n",
+        "DATA_CONVERTER_JOB_PREFIX = \"data_converter\"\n",
+        "DATA_CONVERTER_CONTAINER = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/data-converter:latest\"\n",
+        "DATA_CONVERTER_MACHINE_TYPE = \"n1-highmem-8\"\n",
+        "\n",
+        "\n",
+        "# Training constants.\n",
+        "TRAINING_JOB_PREFIX = \"train\"\n",
+        "TRAIN_CONTAINER_URI = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/tfvision-oss:latest\"\n",
+        "TRAIN_MACHINE_TYPE = \"n1-highmem-16\"\n",
+        "TRAIN_ACCELERATOR_TYPE = \"NVIDIA_TESLA_V100\"\n",
+        "TRAIN_NUM_GPU = 2\n",
+        "TRAIN_DEEPLABV3PLUS_CONFIG = os.path.join(\n",
+        "    CONFIG_DIR, \"deeplabv3plus_resnet101_cityscapes_gpu_multiworker_mirrored.yaml\"\n",
+        ")\n",
+        "\n",
+        "# Evaluation constants.\n",
+        "EVALUATION_METRIC = \"mean_iou\"\n",
+        "\n",
+        "# Export constants.\n",
+        "EXPORT_JOB_PREFIX = \"export\"\n",
+        "EXPORT_CONTAINER_URI = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/tfvision-serving:latest\"\n",
+        "EXPORT_MACHINE_TYPE = \"n1-highmem-8\"\n",
+        "\n",
+        "# Prediction constants.\n",
+        "# You can deploy models with\n",
+        "#   pre-build-dockers: https://cloud.google.com/vertex-ai/docs/predictions/pre-built-containers.\n",
+        "#   and optimized tensorflow runtime dockers: https://cloud.google.com/vertex-ai/docs/predictions/optimized-tensorflow-runtime.\n",
+        "# The example in this notebook uses optimized tensorflow runtime dockers.\n",
+        "# You can adjust accelerator types and machine types to get faster predictions.\n",
+        "PREDICTION_CONTAINER_URI = (\n",
+        "    \"us-docker.pkg.dev/vertex-ai-restricted/prediction/tf_opt-gpu.2-11:latest\"\n",
+        ")\n",
+        "SERVING_CONTAINER_ARGS = [\"--allow_precompilation\", \"--allow_compression\"]\n",
+        "PREDICTION_ACCELERATOR_TYPE = \"NVIDIA_TESLA_T4\"\n",
+        "PREDICTION_MACHINE_TYPE = \"n1-standard-4\"\n",
+        "UPLOAD_JOB_PREFIX = \"upload\"\n",
+        "DEPLOY_JOB_PREFIX = \"deploy\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "ZZFPe_GezXg8"
+      },
+      "source": [
+        "### Define common libraries"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "XcYUGwr-AJGY"
+      },
+      "outputs": [],
+      "source": [
+        "import base64\n",
+        "import json\n",
+        "import os\n",
+        "from datetime import datetime\n",
+        "from io import BytesIO\n",
+        "from typing import Dict, List, Union\n",
+        "\n",
+        "import matplotlib.pyplot as plt\n",
+        "import numpy as np\n",
+        "import tensorflow as tf\n",
+        "import yaml\n",
+        "from google.cloud import aiplatform\n",
+        "from google.protobuf import json_format\n",
+        "from google.protobuf.struct_pb2 import Value\n",
+        "from PIL import Image\n",
+        "\n",
+        "\n",
+        "def get_job_name_with_datetime(prefix: str):\n",
+        "    return prefix + datetime.now().strftime(\"_%Y%m%d_%H%M%S\")\n",
+        "\n",
+        "\n",
+        "def predict_custom_trained_model(\n",
+        "    project: str,\n",
+        "    endpoint_id: str,\n",
+        "    instances: Union[Dict, List[Dict]],\n",
+        "    location: str = \"us-central1\",\n",
+        "    api_endpoint: str = \"us-central1-aiplatform.googleapis.com\",\n",
+        "):\n",
+        "    # The AI Platform services require regional API endpoints.\n",
+        "    client_options = {\"api_endpoint\": api_endpoint}\n",
+        "    # Initialize client that will be used to create and send requests.\n",
+        "    # This client only needs to be created once, and can be reused for multiple requests.\n",
+        "    client = aiplatform.gapic.PredictionServiceClient(client_options=client_options)\n",
+        "    parameters_dict = {}\n",
+        "    parameters = json_format.ParseDict(parameters_dict, Value())\n",
+        "    endpoint = client.endpoint_path(\n",
+        "        project=project, location=location, endpoint=endpoint_id\n",
+        "    )\n",
+        "    response = client.predict(\n",
+        "        endpoint=endpoint, instances=instances, parameters=parameters\n",
+        "    )\n",
+        "    return response.predictions, response.deployed_model_id\n",
+        "\n",
+        "\n",
+        "def load_img(path):\n",
+        "    img = tf.io.read_file(path)\n",
+        "    img = tf.image.decode_jpeg(img, channels=3)\n",
+        "    return Image.fromarray(np.uint8(img)).convert(\"RGB\")\n",
+        "\n",
+        "\n",
+        "def display_image(original_image, category_image_color, score_image_grayscale):\n",
+        "    _, axarr = plt.subplots(1, 3, figsize=(20, 15))\n",
+        "    axarr[0].imshow(original_image)\n",
+        "    axarr[1].imshow(category_image_color)\n",
+        "    axarr[2].imshow(score_image_grayscale.convert(\"RGB\"))\n",
+        "\n",
+        "\n",
+        "def get_prediction_instances(test_filepath, new_width=-1):\n",
+        "    if new_width <= 0:\n",
+        "        with tf.io.gfile.GFile(test_filepath, \"rb\") as input_file:\n",
+        "            encoded_string = base64.b64encode(input_file.read()).decode(\"utf-8\")\n",
+        "    else:\n",
+        "        img = load_img(test_filepath)\n",
+        "        width, height = img.size\n",
+        "        print(\"original input image size: \", width, \" , \", height)\n",
+        "        new_height = int(height * new_width / width)\n",
+        "        new_img = img.resize((new_width, new_height))\n",
+        "        print(\"resized input image size: \", new_width, \" , \", new_height)\n",
+        "        buffered = BytesIO()\n",
+        "        new_img.save(buffered, format=\"JPEG\")\n",
+        "        encoded_string = base64.b64encode(buffered.getvalue()).decode(\"utf-8\")\n",
+        "\n",
+        "    instances = [\n",
+        "        {\n",
+        "            \"encoded_image\": {\"b64\": encoded_string},\n",
+        "        }\n",
+        "    ]\n",
+        "    return instances\n",
+        "\n",
+        "\n",
+        "def get_label_map(label_map_yaml_filepath):\n",
+        "    with tf.io.gfile.GFile(label_map_yaml_filepath, \"rb\") as input_file:\n",
+        "        label_map = yaml.safe_load(input_file.read())\n",
+        "    return label_map\n",
+        "\n",
+        "\n",
+        "def get_best_trial(model_dir, max_trial_count, evaluation_metric):\n",
+        "    best_trial_dir = \"\"\n",
+        "    best_trial_evaluation_results = {}\n",
+        "    best_performance = -1\n",
+        "\n",
+        "    for i in range(max_trial_count):\n",
+        "        current_trial = i + 1\n",
+        "        current_trial_dir = os.path.join(model_dir, \"trial_\" + str(current_trial))\n",
+        "        current_trial_best_ckpt_dir = os.path.join(current_trial_dir, \"best_ckpt\")\n",
+        "        current_trial_best_ckpt_evaluation_filepath = os.path.join(\n",
+        "            current_trial_best_ckpt_dir, \"info.json\"\n",
+        "        )\n",
+        "        with tf.io.gfile.GFile(current_trial_best_ckpt_evaluation_filepath, \"rb\") as f:\n",
+        "            eval_metric_results = json.load(f)\n",
+        "            current_performance = eval_metric_results[evaluation_metric]\n",
+        "            if current_performance > best_performance:\n",
+        "                best_performance = current_performance\n",
+        "                best_trial_dir = current_trial_dir\n",
+        "                best_trial_evaluation_results = eval_metric_results\n",
+        "    return best_trial_dir, best_trial_evaluation_results\n",
+        "\n",
+        "\n",
+        "def create_coco_stuff_label_colormap():\n",
+        "    \"\"\"Creates a label colormap used in COCO-Stuff segmentation benchmark.\n",
+        "\n",
+        "    Returns:\n",
+        "      A colormap for visualizing segmentation results.\n",
+        "    \"\"\"\n",
+        "    return np.asarray(\n",
+        "        [\n",
+        "            [54, 178, 118],\n",
+        "            [0, 85, 178],\n",
+        "            [150, 178, 22],\n",
+        "            [107, 0, 0],\n",
+        "            [0, 0, 89],\n",
+        "            [0, 117, 178],\n",
+        "            [47, 178, 124],\n",
+        "            [178, 116, 0],\n",
+        "            [0, 0, 178],\n",
+        "            [79, 178, 92],\n",
+        "            [134, 0, 0],\n",
+        "            [22, 178, 150],\n",
+        "            [178, 87, 0],\n",
+        "            [178, 146, 0],\n",
+        "            [0, 5, 178],\n",
+        "            [0, 0, 125],\n",
+        "            [0, 53, 178],\n",
+        "            [0, 132, 178],\n",
+        "            [111, 178, 60],\n",
+        "            [178, 131, 0],\n",
+        "            [0, 29, 178],\n",
+        "            [178, 109, 0],\n",
+        "            [178, 35, 0],\n",
+        "            [0, 148, 178],\n",
+        "            [9, 172, 163],\n",
+        "            [0, 0, 178],\n",
+        "            [178, 124, 0],\n",
+        "            [178, 102, 0],\n",
+        "            [0, 156, 175],\n",
+        "            [178, 43, 0],\n",
+        "            [0, 0, 170],\n",
+        "            [178, 94, 0],\n",
+        "            [0, 0, 134],\n",
+        "            [67, 178, 105],\n",
+        "            [99, 178, 73],\n",
+        "            [0, 37, 178],\n",
+        "            [86, 178, 86],\n",
+        "            [15, 178, 156],\n",
+        "            [0, 0, 152],\n",
+        "            [178, 21, 0],\n",
+        "            [0, 124, 178],\n",
+        "            [0, 61, 178],\n",
+        "            [178, 50, 0],\n",
+        "            [0, 109, 178],\n",
+        "            [137, 178, 35],\n",
+        "            [0, 13, 178],\n",
+        "            [0, 101, 178],\n",
+        "            [0, 0, 116],\n",
+        "            [0, 45, 178],\n",
+        "            [41, 178, 131],\n",
+        "            [0, 0, 161],\n",
+        "            [178, 72, 0],\n",
+        "            [0, 0, 143],\n",
+        "            [116, 0, 0],\n",
+        "            [28, 178, 143],\n",
+        "            [170, 6, 0],\n",
+        "            [156, 178, 15],\n",
+        "            [89, 0, 0],\n",
+        "            [143, 178, 28],\n",
+        "            [73, 178, 99],\n",
+        "            [118, 178, 54],\n",
+        "            [92, 178, 79],\n",
+        "            [152, 0, 0],\n",
+        "            [178, 153, 0],\n",
+        "            [98, 0, 0],\n",
+        "            [178, 65, 0],\n",
+        "            [60, 178, 111],\n",
+        "            [169, 175, 3],\n",
+        "            [105, 178, 67],\n",
+        "            [178, 13, 0],\n",
+        "            [163, 178, 9],\n",
+        "            [3, 164, 169],\n",
+        "            [125, 0, 0],\n",
+        "            [175, 168, 0],\n",
+        "            [178, 138, 0],\n",
+        "            [178, 28, 0],\n",
+        "            [35, 178, 137],\n",
+        "            [0, 140, 178],\n",
+        "            [0, 0, 98],\n",
+        "            [131, 178, 41],\n",
+        "            [0, 77, 178],\n",
+        "            [0, 0, 107],\n",
+        "            [0, 93, 178],\n",
+        "            [143, 0, 0],\n",
+        "            [178, 58, 0],\n",
+        "            [161, 0, 0],\n",
+        "            [0, 69, 178],\n",
+        "            [178, 160, 0],\n",
+        "            [178, 80, 0],\n",
+        "            [0, 21, 178],\n",
+        "            [124, 178, 47],\n",
+        "            [255, 214, 0],\n",
+        "        ]\n",
+        "    )\n",
+        "\n",
+        "\n",
+        "def parse_segmentation_prediction(prediction):\n",
+        "    score_bytes = prediction[\"score_bytes\"]\n",
+        "    score_image_grayscale = Image.open(\n",
+        "        BytesIO(base64.b64decode(dict(score_bytes)[\"b64\"]))\n",
+        "    )\n",
+        "    category_bytes = prediction[\"category_bytes\"]\n",
+        "    category_image_grayscale = Image.open(\n",
+        "        BytesIO(base64.b64decode(dict(category_bytes)[\"b64\"]))\n",
+        "    )\n",
+        "\n",
+        "    # Visualize category images.\n",
+        "    color_map = create_coco_stuff_label_colormap()\n",
+        "    category_image_grayscale_np = np.array(category_image_grayscale)\n",
+        "    rendered_image_shape = category_image_grayscale_np.shape + (3,)\n",
+        "    category_image_color_np = np.zeros(rendered_image_shape, dtype=np.uint8)\n",
+        "    unique_labels = np.unique(category_image_grayscale_np)\n",
+        "    for label in unique_labels:\n",
+        "        if label == 0:\n",
+        "            continue\n",
+        "        category_image_color_np[category_image_grayscale_np == label] = color_map[\n",
+        "            label % len(color_map)\n",
+        "        ]\n",
+        "    category_image_color = Image.fromarray(category_image_color_np)\n",
+        "\n",
+        "    return score_image_grayscale, category_image_color"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "RB_xY9ipr7ZU"
+      },
+      "source": [
+        "## Train new models\n",
+        "This section shows how to train new models.\n",
+        "1. Convert input data to training formats\n",
+        "2. Create hyperparameter tuning jobs to train new models\n",
+        "3. Find and export best models\n",
+        "\n",
+        "If you already trained models, please go to the section `Test trained models`."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "zgPO1eR3CYjk"
+      },
+      "source": [
+        "### Prepare input data for training\n",
+        "\n",
+        "Prepare data in the format as described [here](https://cloud.google.com/vertex-ai/docs/image-data/classification/prepare-data), and then convert them to the training formats as below:\n",
+        "\n",
+        "* `input_file_path`: The input file path in coco json formats.\n",
+        "* `split_ratio`: The proportion of data to split into train/validation/test.\n",
+        "* `num_shard`: The number of shards for train/validation/test.\n",
+        "* `output_dir`: The output directory, which will container prepared train/test/validation data."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "IndQ_m6ddUEM"
+      },
+      "outputs": [],
+      "source": [
+        "# This job will convert input data as training format, with given split ratios\n",
+        "# and number of shards on train/test/validation.\n",
+        "data_converter_job_name = get_job_name_with_datetime(\n",
+        "    DATA_CONVERTER_JOB_PREFIX + \"_\" + OBJECTIVE\n",
+        ")\n",
+        "\n",
+        "input_file_path = \"\"  # @param {type:\"string\"}\n",
+        "split_ratio = \"0.8,0.1,0.1\"\n",
+        "num_shard = \"10,10,10\"\n",
+        "data_converter_output_dir = os.path.join(BUCKET_URI, data_converter_job_name)\n",
+        "\n",
+        "\n",
+        "worker_pool_specs = [\n",
+        "    {\n",
+        "        \"machine_spec\": {\n",
+        "            \"machine_type\": DATA_CONVERTER_MACHINE_TYPE,\n",
+        "        },\n",
+        "        \"replica_count\": 1,\n",
+        "        \"container_spec\": {\n",
+        "            \"image_uri\": DATA_CONVERTER_CONTAINER,\n",
+        "            \"command\": [],\n",
+        "            \"args\": [\n",
+        "                \"--input_file_path=%s\" % input_file_path,\n",
+        "                \"--input_file_type=coco_json\",\n",
+        "                \"--objective=%s\" % OBJECTIVE,\n",
+        "                \"--num_shard=%s\" % num_shard,\n",
+        "                \"--split_ratio=%s\" % split_ratio,\n",
+        "                \"--output_dir=%s\" % data_converter_output_dir,\n",
+        "            ],\n",
+        "        },\n",
+        "    }\n",
+        "]\n",
+        "\n",
+        "data_converter_custom_job = aiplatform.CustomJob(\n",
+        "    display_name=data_converter_job_name,\n",
+        "    project=PROJECT_ID,\n",
+        "    worker_pool_specs=worker_pool_specs,\n",
+        "    staging_bucket=STAGING_BUCKET,\n",
+        ")\n",
+        "\n",
+        "data_converter_custom_job.run()\n",
+        "\n",
+        "input_train_data_path = os.path.join(data_converter_output_dir, \"train.tfrecord*\")\n",
+        "input_validation_data_path = os.path.join(data_converter_output_dir, \"val.tfrecord*\")\n",
+        "label_map_path = os.path.join(data_converter_output_dir, \"label_map.yaml\")\n",
+        "print(\"input_train_data_path for training: \", input_train_data_path)\n",
+        "print(\"input_validation_data_path for training: \", input_validation_data_path)\n",
+        "print(\"label_map_path for prediction: \", label_map_path)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "S6dU2IrIqW3H"
+      },
+      "source": [
+        "### Create a Vertex AI custom job with hyperparameter tuning\n",
+        "\n",
+        "You use the Vertex AI SDK to create and run the hyperparameter tuning job with Vertex AI Model Garden Training Dockers."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "aaff6f5be7f6"
+      },
+      "source": [
+        "#### Define the following specifications\n",
+        "* `worker_pool_specs`: Dictionary specifying the machine type and Docker image. This example defines a single node cluster with one `n1-standard-4` machine with two `NVIDIA_TESLA_T4` GPUs.\n",
+        "* `parameter_spec`: Dictionary specifying the parameters to optimize. The dictionary key is the string assigned to the command line argument for each hyperparameter in your training application code, and the dictionary value is the parameter specification. The parameter specification includes the type, min/max values, and scale for the hyperparameter.\n",
+        "* `metric_spec`: Dictionary specifying the metric to optimize. The dictionary key is the `hyperparameter_metric_tag` that you set in your training application code, and the value is the optimization goal."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "um_XKbmpTaHx"
+      },
+      "outputs": [],
+      "source": [
+        "from google.cloud.aiplatform import hyperparameter_tuning as hpt\n",
+        "\n",
+        "label_map = get_label_map(label_map_path)\n",
+        "num_classes = len(label_map[\"label_map\"]) + 1\n",
+        "\n",
+        "# Input train and validation datasets can be found from the section above\n",
+        "# `Convert input data for training`.\n",
+        "# Set prepared datasets if exists.\n",
+        "# input_train_data_path = ''\n",
+        "# input_validation_data_path = ''\n",
+        "\n",
+        "# Refer to https://github.com/tensorflow/models/blob/master/official/vision/MODEL_GARDEN.md\n",
+        "# for more model details.\n",
+        "experiment = \"deeplabv3plus\"  # @param [\"deeplabv3plus\"]\n",
+        "\n",
+        "train_job_name = get_job_name_with_datetime(TRAINING_JOB_PREFIX + \"_\" + OBJECTIVE)\n",
+        "model_dir = os.path.join(BUCKET_URI, train_job_name)\n",
+        "\n",
+        "# The arguments here are mainly for test purposes. Please update them\n",
+        "# to get better performances.\n",
+        "experiment_container_args_dict = {\n",
+        "    # deeplabv3plus experiment args.\n",
+        "    \"deeplabv3plus\": {\n",
+        "        \"experiment\": \"seg_deeplabv3plus_pascal\",\n",
+        "        \"config_file\": TRAIN_DEEPLABV3PLUS_CONFIG,\n",
+        "        \"input_train_data_path\": input_train_data_path,\n",
+        "        \"input_validation_data_path\": input_validation_data_path,\n",
+        "        \"objective\": OBJECTIVE,\n",
+        "        \"model_dir\": model_dir,\n",
+        "        \"num_classes\": num_classes,\n",
+        "        \"global_batch_size\": 2,\n",
+        "        \"prefetch_buffer_size\": 12,\n",
+        "        \"train_steps\": 500,\n",
+        "        \"output_size\": \"1024,2048\",\n",
+        "    }\n",
+        "}\n",
+        "\n",
+        "worker_pool_specs = [\n",
+        "    {\n",
+        "        \"machine_spec\": {\n",
+        "            \"machine_type\": TRAIN_MACHINE_TYPE,\n",
+        "            \"accelerator_type\": TRAIN_ACCELERATOR_TYPE,\n",
+        "            \"accelerator_count\": TRAIN_NUM_GPU,\n",
+        "        },\n",
+        "        \"replica_count\": 1,\n",
+        "        \"container_spec\": {\n",
+        "            \"image_uri\": TRAIN_CONTAINER_URI,\n",
+        "            \"args\": [\n",
+        "                \"--mode=train_and_eval\",\n",
+        "            ]\n",
+        "            + [\n",
+        "                \"--{}={}\".format(k, v)\n",
+        "                for k, v in experiment_container_args_dict[experiment].items()\n",
+        "            ],\n",
+        "        },\n",
+        "    },\n",
+        "]\n",
+        "\n",
+        "metric_spec = {\"model_performance\": \"maximize\"}\n",
+        "\n",
+        "LEARNING_RATES = [0.001]\n",
+        "# Models will be trained with each learning rate separately and max trial count is the number of learning rates.\n",
+        "MAX_TRIAL_COUNT = len(LEARNING_RATES)\n",
+        "parameter_spec = {\n",
+        "    \"learning_rate\": hpt.DiscreteParameterSpec(values=LEARNING_RATES, scale=\"linear\"),\n",
+        "}\n",
+        "\n",
+        "print(worker_pool_specs, metric_spec, parameter_spec)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "HwcCjwlBTQIz"
+      },
+      "source": [
+        "#### Run hyperparameter tuning jobs\n",
+        "* `max_trial_count`: Sets an upper bound on the number of trials the service will run. The recommended practice is to start with a smaller number of trials and get a sense of how impactful your chosen hyperparameters are before scaling up.\n",
+        "\n",
+        "* `parallel_trial_count`:  If you use parallel trials, the service provisions multiple training processing clusters. The worker pool spec that you specify when creating the job is used for each individual training cluster.  Increasing the number of parallel trials reduces the amount of time the hyperparameter tuning job takes to run; however, it can reduce the effectiveness of the job overall. This is because the default tuning strategy uses results of previous trials to inform the assignment of values in subsequent trials.\n",
+        "\n",
+        "* `search_algorithm`: The available search algorithms are grid, random, or default (None). The default option applies Bayesian optimization to search the space of possible hyperparameter values and is the recommended algorithm.\n",
+        "\n",
+        "Click on the generated link in the output to see your run in the Cloud Console."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "aec22792ee84"
+      },
+      "outputs": [],
+      "source": [
+        "train_custom_job = aiplatform.CustomJob(\n",
+        "    display_name=train_job_name,\n",
+        "    project=PROJECT_ID,\n",
+        "    worker_pool_specs=worker_pool_specs,\n",
+        "    staging_bucket=STAGING_BUCKET,\n",
+        ")\n",
+        "\n",
+        "train_hpt_job = aiplatform.HyperparameterTuningJob(\n",
+        "    display_name=train_job_name,\n",
+        "    custom_job=train_custom_job,\n",
+        "    metric_spec=metric_spec,\n",
+        "    parameter_spec=parameter_spec,\n",
+        "    max_trial_count=MAX_TRIAL_COUNT,\n",
+        "    parallel_trial_count=1,\n",
+        "    project=PROJECT_ID,\n",
+        "    search_algorithm=None,\n",
+        ")\n",
+        "\n",
+        "train_hpt_job.run()\n",
+        "\n",
+        "print(\"experiment is: \", experiment)\n",
+        "print(\"model_dir is: \", model_dir)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "mV-Djz-frBni"
+      },
+      "source": [
+        "### Export best models as TF Saved Model format"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "09Rz1AYspK19"
+      },
+      "outputs": [],
+      "source": [
+        "# This job will export models from TF checkpoints to TF saved model format.\n",
+        "from google.cloud import aiplatform\n",
+        "\n",
+        "# model_dir is from the section above.\n",
+        "best_trial_dir, best_trial_evaluation_results = get_best_trial(\n",
+        "    model_dir, MAX_TRIAL_COUNT, EVALUATION_METRIC\n",
+        ")\n",
+        "print(\"best_trial_dir: \", best_trial_dir)\n",
+        "print(\"best_trial_evaluation_results: \", best_trial_evaluation_results)\n",
+        "\n",
+        "worker_pool_specs = [\n",
+        "    {\n",
+        "        \"machine_spec\": {\n",
+        "            \"machine_type\": EXPORT_MACHINE_TYPE,\n",
+        "        },\n",
+        "        \"replica_count\": 1,\n",
+        "        \"container_spec\": {\n",
+        "            \"image_uri\": EXPORT_CONTAINER_URI,\n",
+        "            \"command\": [],\n",
+        "            \"args\": [\n",
+        "                \"--objective=%s\" % OBJECTIVE,\n",
+        "                \"--experiment=%s\"\n",
+        "                % experiment_container_args_dict[experiment][\"experiment\"],\n",
+        "                \"--config_file=%s/params.yaml\" % best_trial_dir,\n",
+        "                \"--checkpoint_path=%s/best_ckpt\" % best_trial_dir,\n",
+        "                \"--export_dir=%s/best_model\" % model_dir,\n",
+        "                \"--input_image_size=%s\"\n",
+        "                % experiment_container_args_dict[experiment][\"output_size\"],\n",
+        "            ],\n",
+        "        },\n",
+        "    }\n",
+        "]\n",
+        "\n",
+        "model_export_name = get_job_name_with_datetime(EXPORT_JOB_PREFIX + \"_\" + OBJECTIVE)\n",
+        "model_export_custom_job = aiplatform.CustomJob(\n",
+        "    display_name=model_export_name,\n",
+        "    project=PROJECT_ID,\n",
+        "    worker_pool_specs=worker_pool_specs,\n",
+        "    staging_bucket=STAGING_BUCKET,\n",
+        ")\n",
+        "\n",
+        "\n",
+        "model_export_custom_job.run()\n",
+        "\n",
+        "print(\"best model is saved to: \", os.path.join(model_dir, \"best_model\"))"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "g0BGaofgsMsy"
+      },
+      "source": [
+        "## Test trained models\n",
+        "This section shows how to test with trained models.\n",
+        "1. Upload and deploy models\n",
+        "2. Run predictions"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "NYuQowyZEtxK"
+      },
+      "outputs": [],
+      "source": [
+        "# @title Upload and deploy models\n",
+        "# model_dir is from the section above.\n",
+        "trained_model_dir = os.path.join(model_dir, \"best_model/saved_model\")\n",
+        "\n",
+        "upload_job_name = get_job_name_with_datetime(UPLOAD_JOB_PREFIX + \"_\" + OBJECTIVE)\n",
+        "\n",
+        "model = aiplatform.Model.upload(\n",
+        "    display_name=upload_job_name,\n",
+        "    artifact_uri=trained_model_dir,\n",
+        "    serving_container_image_uri=PREDICTION_CONTAINER_URI,\n",
+        "    serving_container_args=SERVING_CONTAINER_ARGS,\n",
+        ")\n",
+        "\n",
+        "model.wait()\n",
+        "\n",
+        "print(\"The uploaded model name is: \", upload_job_name)\n",
+        "\n",
+        "deploy_model_name = get_job_name_with_datetime(DEPLOY_JOB_PREFIX + \"_\" + OBJECTIVE)\n",
+        "print(\"The deployed job name is: \", deploy_model_name)\n",
+        "\n",
+        "endpoint = model.deploy(\n",
+        "    deployed_model_display_name=deploy_model_name,\n",
+        "    machine_type=PREDICTION_MACHINE_TYPE,\n",
+        "    traffic_split={\"0\": 100},\n",
+        "    accelerator_type=PREDICTION_ACCELERATOR_TYPE,\n",
+        "    accelerator_count=1,\n",
+        "    min_replica_count=1,\n",
+        "    max_replica_count=1,\n",
+        ")\n",
+        "\n",
+        "endpoint_id = endpoint.name\n",
+        "print(\"endpoint id is: \", endpoint_id)"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "vbIW9me1F2RY"
+      },
+      "outputs": [],
+      "source": [
+        "# @title Run predictions\n",
+        "# endpoint_id was generated in the section above (`Upload and deploy models`).\n",
+        "endpoint_id = endpoint.name\n",
+        "\n",
+        "# The test image file path.\n",
+        "test_filepath = \"\"  # @param {type:\"string\"}\n",
+        "score_threshold = 0.5  # @param {type:\"number\"}\n",
+        "# If the input image is too large, we will resize it for prediction.\n",
+        "instances = get_prediction_instances(test_filepath, new_width=1000)\n",
+        "\n",
+        "# The label map file was generated from the section above (`Convert input data for training`).\n",
+        "label_map = get_label_map(label_map_path)[\"label_map\"]\n",
+        "\n",
+        "predictions, _ = predict_custom_trained_model(\n",
+        "    project=PROJECT_ID, location=REGION, endpoint_id=endpoint_id, instances=instances\n",
+        ")\n",
+        "\n",
+        "score_image_grayscale, category_image_color = parse_segmentation_prediction(\n",
+        "    dict(predictions[0])\n",
+        ")\n",
+        "display_image(\n",
+        "    load_img(test_filepath), category_image_color, score_image_grayscale.convert(\"RGB\")\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "kkH2nrpdp4sp"
+      },
+      "source": [
+        "## Clean up"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "Ax6vQVZhp9pR"
+      },
+      "outputs": [],
+      "source": [
+        "# Undeploy model and delete endpoint.\n",
+        "endpoint.delete(force=True)\n",
+        "# Delete models.\n",
+        "model.delete()\n",
+        "# Delete custom and hpt jobs.\n",
+        "if data_converter_custom_job.list(filter=f'display_name=\"{data_converter_job_name}\"'):\n",
+        "    data_converter_custom_job.delete()\n",
+        "if train_hpt_job.list(filter=f'display_name=\"{train_job_name}\"'):\n",
+        "    train_hpt_job.delete()\n",
+        "if model_export_custom_job.list(filter=f'display_name=\"{model_export_name}\"'):\n",
+        "    model_export_custom_job.delete()"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "name": "model_garden_tfvision_image_segmentation.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/community/model_registry/vertex_ai_model_registry_bqml_custom_model_versioning.ipynb

@@ -198,7 +198,7 @@
         "if IS_WORKBENCH_NOTEBOOK:\n",
         "    USER_FLAG = \"--user\"\n",
         "\n",
-        "! pip3 install --upgrade tensorflow google-cloud-bigquery google-cloud-aiplatform {USER_FLAG} -q --no-warn-conflicts"
+        "! pip3 install --upgrade tensorflow google-cloud-bigquery google-cloud-aiplatform \"shapely<2\" {USER_FLAG} -q --no-warn-conflicts"
       ]
     },
     {

notebooks/community/pipelines/google_cloud_pipeline_components_bqml_pipeline_anomaly_detection.ipynb

@@ -72,7 +72,7 @@
       "source": [
         "## Overview\n",
         "\n",
-        "Anomaly detection is the identification of rare obesrvations which deviate significantly from the data using ML. Anomaly detection can be done in many ways. Supervised, unsupervised, graph-based. It is particularly important for certain industries like telecommunications, manufacturing, and financial services.\n",
+        "Anomaly detection is the identification of rare observations which deviate significantly from the data using ML. Anomaly detection can be done in many ways. Supervised, unsupervised, graph-based. It is particularly important for certain industries like telecommunications, manufacturing, and financial services.\n",
         "\n",
         "For instance, in a manufacturing scenario, you may collect some sensor data to predict the number remaining cycles before engine failure (TTF). In this way, you can take actionable decisions about maintenance planning."
       ]
@@ -397,13 +397,12 @@
       "source": [
         "import os\n",
         "\n",
-        "SRC_PATH = \"src\"\n",
         "KFP_COMPONENTS_PATH = \"components\"\n",
         "PIPELINES_PATH = \"pipelines\"\n",
         "TRAIN_PIPELINES_PATH = os.path.join(PIPELINES_PATH, \"train_pipelines\")\n",
         "TEST_PIPELINES_PATH = os.path.join(PIPELINES_PATH, \"test_pipelines\")\n",
         "\n",
-        "! mkdir -m 777 -p {SRC_PATH} {KFP_COMPONENTS_PATH} {TRAIN_PIPELINES_PATH} {TEST_PIPELINES_PATH}"
+        "! mkdir -m 777 -p {KFP_COMPONENTS_PATH} {TRAIN_PIPELINES_PATH} {TEST_PIPELINES_PATH}"
       ]
     },
     {
@@ -425,14 +424,12 @@
       },
       "outputs": [],
       "source": [
-        "from urllib.parse import urlparse\n",
-        "\n",
         "PUBLIC_DATA_URI = (\n",
-        "    \" gs://cloud-samples-data/vertex-ai/pipeline-deployment/datasets/turbofan_anomaly\"\n",
+        "    \"gs://cloud-samples-data/vertex-ai/pipeline-deployment/datasets/turbofan_anomaly\"\n",
         ")\n",
-        "GCS_TRAIN_URI = urlparse(PUBLIC_DATA_URI)._replace(path=\"train_FD001.csv\").geturl()\n",
-        "GCS_TEST_URI = urlparse(PUBLIC_DATA_URI)._replace(path=\"test_FD001.csv\").geturl()\n",
-        "GCS_LABELS_URI = urlparse(PUBLIC_DATA_URI)._replace(path=\"RUL_FD001.csv\").geturl()"
+        "GCS_TRAIN_URI = f\"{PUBLIC_DATA_URI}/train_FD001.csv\"\n",
+        "GCS_TEST_URI = f\"{PUBLIC_DATA_URI}/test_FD001.csv\"\n",
+        "GCS_LABELS_URI = f\"{PUBLIC_DATA_URI}/RUL_FD001.csv\""
       ]
     },
     {
@@ -1478,7 +1475,6 @@
         "# Remove local resorces\n",
         "delete_local_resources = False\n",
         "if delete_local_resources:\n",
-        "    ! rm -rf {SRC_PATH}\n",
         "    ! rm -rf {KFP_COMPONENTS_PATH}\n",
         "    ! rm -rf {TRAIN_PIPELINES_PATH}\n",
         "    ! rm -rf {TEST_PIPELINES_PATH}"

notebooks/community/pipelines/google_cloud_pipeline_components_bqml_pipeline_demand_forecasting.ipynb

@@ -54,18 +54,20 @@
     {
       "cell_type": "markdown",
       "metadata": {
-        "id": "tvgnzT1CKxrO"
+        "id": "239ba71252d3"
       },
       "source": [
         "## Overview\n",
         "\n",
-        "This notebook shows how to use `Vertex AI Pipelines` and `BigQuery ML pipeline components` to train and evaluate a demand forecasting model.\n",
-        "\n",
-        "### Dataset\n",
-        "\n",
-        "The dataset is a modified version of the dataset in [Build and visualize demand forecast predictions using Datastream, Dataflow, BigQuery ML, and Looker\n",
-        "](https://cloud.google.com/architecture/build-visualize-demand-forecast-prediction-datastream-dataflow-bigqueryml-looker) solution architecture\n",
-        "\n",
+        "This notebook shows how to use `Vertex AI Pipelines` and `BigQuery ML pipeline components` to train and evaluate a demand forecasting model."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "25c28706c23e"
+      },
+      "source": [
         "### Objective\n",
         "\n",
         "In this tutorial, you learn how to train and evaluate a BigQuery ML model using Vertex AI Pipelines and BigQuery ML pipeline components. \n",
@@ -87,8 +89,27 @@
         "  - Generate the ARIMA Plus forecasts\n",
         "  - Generate the ARIMA PLUS forecast explainations\n",
         "- Compile the pipeline.\n",
-        "- Execute the pipeline.\n",
+        "- Execute the pipeline."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "586acfa9b502"
+      },
+      "source": [
+        "### Dataset\n",
         "\n",
+        "The dataset is a modified version of the dataset in [Build and visualize demand forecast predictions using Datastream, Dataflow, BigQuery ML, and Looker\n",
+        "](https://cloud.google.com/architecture/build-visualize-demand-forecast-prediction-datastream-dataflow-bigqueryml-looker) solution architecture\n"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "tvgnzT1CKxrO"
+      },
+      "source": [
         "### Costs \n",
         "\n",
         "This tutorial uses billable components of Google Cloud:\n",
@@ -352,9 +373,8 @@
         "id": "06571eb4063b"
       },
       "source": [
-        "#### Timestamp\n",
-        "\n",
-        "If you are in a live tutorial session, you might be using a shared test account or project. To avoid name collisions between users on resources created, you create a timestamp for each instance session, and append it onto the name of resources you create in this tutorial."
+        "#### UUID\n",
+        "If you are in a live tutorial session, you might be using a shared test account or project. To avoid name collisions between users on resources created, you create a uuid for each instance session, and append it onto the name of resources you create in this tutorial."
       ]
     },
     {
@@ -365,9 +385,16 @@
       },
       "outputs": [],
       "source": [
-        "from datetime import datetime\n",
+        "import random\n",
+        "import string\n",
         "\n",
-        "TIMESTAMP = datetime.now().strftime(\"%Y%m%d%H%M%S\")"
+        "\n",
+        "# Generate a uuid of a specifed length(default=8)\n",
+        "def generate_uuid(length: int = 8) -> str:\n",
+        "    return \"\".join(random.choices(string.ascii_lowercase + string.digits, k=length))\n",
+        "\n",
+        "\n",
+        "UUID = generate_uuid()"
       ]
     },
     {
@@ -485,7 +512,7 @@
       "outputs": [],
       "source": [
         "if BUCKET_NAME == \"\" or BUCKET_NAME is None or BUCKET_NAME == \"[your-bucket-name]\":\n",
-        "    BUCKET_NAME = PROJECT_ID + \"-aip-\" + TIMESTAMP\n",
+        "    BUCKET_NAME = PROJECT_ID + \"-aip-\" + UUID\n",
         "    BUCKET_URI = f\"gs://{BUCKET_NAME}\""
       ]
     },
@@ -706,6 +733,7 @@
         "KFP_COMPONENTS_PATH = \"components\"\n",
         "PIPELINES_PATH = \"pipelines\"\n",
         "\n",
+        "! mkdir -m 777 -p {DATA_PATH}\n",
         "! mkdir -m 777 -p {KFP_COMPONENTS_PATH}\n",
         "! mkdir -m 777 -p {PIPELINES_PATH}"
       ]
@@ -771,7 +799,7 @@
         "  --location={LOCATION} \\\n",
         "  --source_format=CSV \\\n",
         "  --skip_leading_rows=1\\\n",
-        "  fast_fresh.orders_{TIMESTAMP} \\\n",
+        "  fast_fresh.orders_{UUID} \\\n",
         "  {RAW_DATA_URI} \\\n",
         "  time_of_sale:DATETIME,order_id:INTEGER,product_name:STRING,price:NUMERIC,quantity:NUMERIC,payment_method:STRING,store_id:INTEGER,user_id:INTEGER"
       ]
@@ -782,7 +810,7 @@
         "id": "ZrgOD30o7HcL"
       },
       "source": [
-        "## BQML Training Formalization\n",
+        "## BigQuery ML Training Formalization\n",
         "\n",
         "In the next cells, you build the components and pipeline to train and evaluate the BQML demand forecasting model."
       ]
@@ -820,13 +848,13 @@
         "BQ_EVALUATE_MODEL_TABLE_PREFIX = \"orders_arima_model_evaluate\"\n",
         "BQ_FORECAST_TABLE_PREFIX = \"orders_arima_forecast\"\n",
         "BQ_EXPLAIN_FORECAST_TABLE_PREFIX = \"orders_arima_explain_forecast\"\n",
-        "BQ_ORDERS_TABLE = f\"{BQ_ORDERS_TABLE_PREFIX}_{TIMESTAMP}\"\n",
-        "BQ_TRAINING_TABLE = f\"{BQ_TRAINING_TABLE_PREFIX}_{TIMESTAMP}\"\n",
-        "BQ_MODEL_TABLE = f\"{BQ_MODEL_TABLE_PREFIX}_{TIMESTAMP}\"\n",
-        "BQ_EVALUATE_TS_TABLE = f\"{BQ_EVALUATE_TS_TABLE_PREFIX}_{TIMESTAMP}\"\n",
-        "BQ_EVALUATE_MODEL_TABLE = f\"{BQ_EVALUATE_MODEL_TABLE_PREFIX}_{TIMESTAMP}\"\n",
-        "BQ_FORECAST_TABLE = f\"{BQ_FORECAST_TABLE_PREFIX}_{TIMESTAMP}\"\n",
-        "BQ_EXPLAIN_FORECAST_TABLE = f\"{BQ_EXPLAIN_FORECAST_TABLE_PREFIX}_{TIMESTAMP}\"\n",
+        "BQ_ORDERS_TABLE = f\"{BQ_ORDERS_TABLE_PREFIX}_{UUID}\"\n",
+        "BQ_TRAINING_TABLE = f\"{BQ_TRAINING_TABLE_PREFIX}_{UUID}\"\n",
+        "BQ_MODEL_TABLE = f\"{BQ_MODEL_TABLE_PREFIX}_{UUID}\"\n",
+        "BQ_EVALUATE_TS_TABLE = f\"{BQ_EVALUATE_TS_TABLE_PREFIX}_{UUID}\"\n",
+        "BQ_EVALUATE_MODEL_TABLE = f\"{BQ_EVALUATE_MODEL_TABLE_PREFIX}_{UUID}\"\n",
+        "BQ_FORECAST_TABLE = f\"{BQ_FORECAST_TABLE_PREFIX}_{UUID}\"\n",
+        "BQ_EXPLAIN_FORECAST_TABLE = f\"{BQ_EXPLAIN_FORECAST_TABLE_PREFIX}_{UUID}\"\n",
         "\n",
         "BQ_TRAIN_CONFIGURATION = {\n",
         "    \"destinationTable\": {\n",
@@ -1022,7 +1050,7 @@
         "id": "pcSL1FHk69KT"
       },
       "source": [
-        "### Build the BQML training pipeline\n",
+        "### Build the BigQuery ML training pipeline\n",
         "\n",
         "Define your workflow using Kubeflow Pipelines DSL package. \n",
         "\n",
@@ -1094,8 +1122,8 @@
         "        location=location,\n",
         "    ).set_display_name(\"get train data\")\n",
         "\n",
-        "    # Train the ARIMA PLUS model\n",
-        "    bq_arima_model_op = (\n",
+        "    # Run an ARIMA PLUS experiment\n",
+        "    bq_arima_model_exp_op = (\n",
         "        BigqueryCreateModelJobOp(\n",
         "            query=f\"\"\"\n",
         "        -- create model table\n",
@@ -1104,10 +1132,7 @@
         "        MODEL_TYPE = \\'ARIMA_PLUS\\',\n",
         "        TIME_SERIES_TIMESTAMP_COL = \\'hourly_timestamp\\',\n",
         "        TIME_SERIES_DATA_COL = \\'total_sold\\',\n",
-        "        TIME_SERIES_ID_COL = [\\'product_name\\'],\n",
-        "        MODEL_REGISTRY = \\'vertex_ai\\',\n",
-        "        VERTEX_AI_MODEL_ID = \\'order_demand_forecasting\\',\n",
-        "        VERTEX_AI_MODEL_VERSION_ALIASES = [\\'staging\\']\n",
+        "        TIME_SERIES_ID_COL = [\\'product_name\\']\n",
         "        ) AS\n",
         "        SELECT\n",
         "          hourly_timestamp,\n",
@@ -1119,7 +1144,7 @@
         "            project=project,\n",
         "            location=location,\n",
         "        )\n",
-        "        .set_display_name(\"train arima plus model\")\n",
+        "        .set_display_name(\"run arima+ model experiment\")\n",
         "        .after(create_training_dataset_op)\n",
         "    )\n",
         "\n",
@@ -1128,12 +1153,12 @@
         "        BigqueryMLArimaEvaluateJobOp(\n",
         "            project=project,\n",
         "            location=location,\n",
-        "            model=bq_arima_model_op.outputs[\"model\"],\n",
+        "            model=bq_arima_model_exp_op.outputs[\"model\"],\n",
         "            show_all_candidate_models=False,\n",
         "            job_configuration_query=bq_evaluate_time_series_configuration,\n",
         "        )\n",
         "        .set_display_name(\"evaluate arima plus time series\")\n",
-        "        .after(bq_arima_model_op)\n",
+        "        .after(bq_arima_model_exp_op)\n",
         "    )\n",
         "\n",
         "    # Evaluate ARIMA Plus model\n",
@@ -1141,12 +1166,12 @@
         "        BigqueryEvaluateModelJobOp(\n",
         "            project=project,\n",
         "            location=location,\n",
-        "            model=bq_arima_model_op.outputs[\"model\"],\n",
+        "            model=bq_arima_model_exp_op.outputs[\"model\"],\n",
         "            query_statement=f\"\"\"SELECT * FROM `{project}.{bq_dataset}.{bq_training_table}` WHERE split='TEST'\"\"\",\n",
         "            job_configuration_query=bq_evaluate_model_configuration,\n",
         "        )\n",
         "        .set_display_name(\"evaluate arima plus model\")\n",
-        "        .after(bq_arima_model_op)\n",
+        "        .after(bq_arima_model_exp_op)\n",
         "    )\n",
         "\n",
         "    # Plot model metrics\n",
@@ -1164,6 +1189,34 @@
         "        < PERF_THRESHOLD,\n",
         "        name=\"avg. mae good\",\n",
         "    ):\n",
+        "        # Train the ARIMA PLUS model\n",
+        "        bq_arima_model_op = (\n",
+        "            BigqueryCreateModelJobOp(\n",
+        "                query=f\"\"\"\n",
+        "        -- create model table\n",
+        "        CREATE OR REPLACE MODEL `{project}.{bq_dataset}.{bq_model_table}`\n",
+        "        OPTIONS(\n",
+        "        MODEL_TYPE = \\'ARIMA_PLUS\\',\n",
+        "        TIME_SERIES_TIMESTAMP_COL = \\'hourly_timestamp\\',\n",
+        "        TIME_SERIES_DATA_COL = \\'total_sold\\',\n",
+        "        TIME_SERIES_ID_COL = [\\'product_name\\'],\n",
+        "        MODEL_REGISTRY = \\'vertex_ai\\',\n",
+        "        VERTEX_AI_MODEL_ID = \\'order_demand_forecasting\\',\n",
+        "        VERTEX_AI_MODEL_VERSION_ALIASES = [\\'staging\\']\n",
+        "        ) AS\n",
+        "        SELECT\n",
+        "          DATETIME_TRUNC(time_of_sale, HOUR) as hourly_timestamp,\n",
+        "          product_name,\n",
+        "          SUM(quantity) AS total_sold,\n",
+        "          FROM `{project}.{bq_dataset}.{bq_orders_table}`\n",
+        "        GROUP BY hourly_timestamp, product_name;\n",
+        "        \"\"\",\n",
+        "                project=project,\n",
+        "                location=location,\n",
+        "            )\n",
+        "            .set_display_name(\"train arima+ model\")\n",
+        "            .after(get_evaluation_model_metrics_op)\n",
+        "        )\n",
         "\n",
         "        # Generate the ARIMA PLUS forecasts\n",
         "        bq_arima_forecast_op = (\n",
@@ -1224,7 +1277,7 @@
       "source": [
         "### Execute your pipeline\n",
         "\n",
-        "Next, you execute the pipeline. It takes the following parameters which we set as default:\n",
+        "Next, we execute the pipeline. It takes the following parameters which we set as default:\n",
         "\n",
         "- `bq_dataset`: The BigQuery dataset to train on.\n",
         "- `bq_orders_table` : The BigQuery table of raw data.\n",
@@ -1266,7 +1319,7 @@
       "source": [
         "### View BigQuery ML training pipeline results\n",
         "\n",
-        "Finally, you will view the artifact outputs of each task in the pipeline."
+        "Finally, you view the artifact outputs of each task in the pipeline."
       ]
     },
     {
@@ -1342,8 +1395,8 @@
         "print(\"bigquery-ml-arima-evaluate-job\")\n",
         "artifacts = print_pipeline_output(bqml_pipeline, \"bigquery-ml-arima-evaluate-job\")\n",
         "print(\"\\n\\n\")\n",
-        "print(\"get-model-evaluation-metrics\")\n",
-        "artifacts = print_pipeline_output(bqml_pipeline, \"get-model-evaluation-metrics\")\n",
+        "print(\"bigquery-evaluate-model-job\")\n",
+        "artifacts = print_pipeline_output(bqml_pipeline, \"bigquery-evaluate-model-job\")\n",
         "print(\"\\n\\n\")\n",
         "print(\"bigquery-forecast-model-job\")\n",
         "artifacts = print_pipeline_output(bqml_pipeline, \"bigquery-forecast-model-job\")\n",
@@ -1407,7 +1460,8 @@
         "\n",
         "# Remove local resorces\n",
         "! rm -rf {KFP_COMPONENTS_PATH}\n",
-        "! rm -rf {PIPELINES_PATH}"
+        "! rm -rf {PIPELINES_PATH}\n",
+        "! rm -rf {DATA_PATH}"
       ]
     }
   ],

notebooks/community/pipelines/google_cloud_pipeline_components_cloud_natural_language_pipeline.ipynb

@@ -0,0 +1,870 @@
+{
+  "cells": [
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "view-in-github"
+      },
+      "source": [
+        "<a href=\"https://colab.research.google.com/github/Narwhalprime/vertex-ai-samples/blob/main/notebooks/community/pipelines/google_cloud_pipeline_components_cloud_natural_language_pipeline.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "1142fd18"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2022 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "BwO30Ag12YcB"
+      },
+      "source": [
+        "# Vertex Pipelines: Cloud Natural Language model training pipeline\n",
+        "<table align=\"left\">\n",
+        "\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/pipelines/google_cloud_pipeline_components_cloud_natural_language_pipeline.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/pipelines/google_cloud_pipeline_components_cloud_natural_language_pipeline.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>\n",
+        "    <a href=\"https://console.cloud.google.com/ai/platform/notebooks/deploy-notebook?download_url=https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/natural_language/cloud_natural_language_pipeline.ipynb\">\n",
+        "      <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "      Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "tvgnzT1CKxrO"
+      },
+      "source": [
+        "## Overview\n",
+        "This notebook shows how to use [Google Cloud Pipeline Components SDK](https://cloud.google.com/vertex-ai/docs/pipelines/components-introduction) and additional components in this directory to run a machine learning pipeline in [Vertex AI Pipelines](https://cloud.google.com/vertex-ai/docs/pipelines/introduction) to train a TensorFlow text classification model.\n",
+        "\n",
+        "In this pipeline, the model training Docker image utilizes [TFHub](https://tfhub.dev/) models to perform state-of-the-art text classification training. The image is pre-built and ready to use, so no additional Docker setup is required."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d975e698c9a4"
+      },
+      "source": [
+        "### Objective\n",
+        "\n",
+        "In this tutorial, you learn how to construct an end-to-end training pipeine within Vertex AI pipelines that ingests a dataset, trains a text classification model on it, and outputs evaluation metrics.\n",
+        "\n",
+        "This tutorial uses the following Google Cloud ML services and resources:\n",
+        "\n",
+        "- Vertex AI Pipelines\n",
+        "- Vertex AI Datasets\n",
+        "\n",
+        "The steps performed include:\n",
+        "\n",
+        "- Define Kubeflow pipeline components\n",
+        "- Setup Kubeflow pipeline\n",
+        "- Run pipeline on Vertex AI"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "08d289fa873f"
+      },
+      "source": [
+        "## Dataset\n",
+        "\n",
+        "This notebook requires that the user has two datasets exported from Vertex AI [managed datasets](https://cloud.google.com/vertex-ai/docs/training/using-managed-datasets): one with train and validation data splits, and the other with test data used for evaluation. Please ensure no data is shared between the two datasets (in particular, no evaluation data should be part of the train or validation splits). To export a Vertex AI dataset, please follow the following public docs:\n",
+        "* [Preparing data](https://cloud.google.com/vertex-ai/docs/text-data/classification/prepare-data)\n",
+        "* [Creating a Vertex AI dataset](https://cloud.google.com/vertex-ai/docs/text-data/classification/create-dataset) from the above data\n",
+        "* [Exporting dataset and its annotations](https://cloud.google.com/vertex-ai/docs/datasets/export-metadata-annotations); ensure the resulting export is located in a Google Cloud Storage (GCS) bucket you own. You may need to manually separate the test split data into its own file."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "aed92deeb4a0"
+      },
+      "source": [
+        "## Costs\n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI\n",
+        "* Cloud Storage\n",
+        "\n",
+        "Learn about [Vertex AI\n",
+        "pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage\n",
+        "pricing](https://cloud.google.com/storage/pricing), and use the [Pricing\n",
+        "Calculator](https://cloud.google.com/products/calculator/)\n",
+        "to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "setup_local"
+      },
+      "source": [
+        "## Setup\n",
+        "\n",
+        "If you are using Colab or Google Vertex AI Workbench Notebooks, your environment already meets all the requirements to run this notebook. You can skip this step.\n",
+        "\n",
+        "***NOTE***: This notebook has been tested in the following environment:\n",
+        "\n",
+        "* Python version = 3.8\n",
+        "\n",
+        "Otherwise, make sure your environment meets this notebook's requirements. You need the following:\n",
+        "\n",
+        "- The Cloud Storage SDK\n",
+        "- Python 3\n",
+        "- virtualenv\n",
+        "- Jupyter notebook running in a virtual environment with Python 3\n",
+        "\n",
+        "The Cloud Storage guide to [Setting up a Python development environment](https://cloud.google.com/python/setup) and the [Jupyter installation guide](https://jupyter.org/install) provide detailed instructions for meeting these requirements. The following steps provide a condensed set of instructions:\n",
+        "\n",
+        "1. [Install and initialize the SDK](https://cloud.google.com/sdk/docs/).\n",
+        "\n",
+        "2. [Install Python 3](https://cloud.google.com/python/setup#installing_python).\n",
+        "\n",
+        "3. [Install virtualenv](https://cloud.google.com/python/setup#installing_and_using_virtualenv) and create a virtual environment that uses Python 3. Activate the virtual environment.\n",
+        "\n",
+        "4. Activate that environment and run `pip3 install Jupyter` in a terminal shell to install Jupyter.\n",
+        "\n",
+        "5. Run `jupyter notebook` on the command line in a terminal shell to launch Jupyter.\n",
+        "\n",
+        "6. Open this notebook in the Jupyter Notebook Dashboard.\n"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "568d5c16"
+      },
+      "source": [
+        "### Install additional packages\n",
+        "\n",
+        "Run the following commands to setup the packages for this notebook. Note that the last code snippet in this section restarts your kernel in order to load the installs properly, so when initalizing this notebook from scratch, it is recommended to run up to that cell, then afterwards you may start running the cell after that."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "dac98aac"
+      },
+      "outputs": [],
+      "source": [
+        "# Install using pip3\n",
+        "!pip3 install -U tensorflow google-cloud-pipeline-components google-cloud-aiplatform kfp==1.8.16 \"shapely<2\" -q"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "alRWYgYTdz7P"
+      },
+      "outputs": [],
+      "source": [
+        "# Version check\n",
+        "# This has been tested with KFP 1.8.16\n",
+        "! python3 -c \"import kfp; print('KFP SDK version: {}'.format(kfp.__version__))\"\n",
+        "! python3 -c \"import google_cloud_pipeline_components; print('google_cloud_pipeline_components version: {}'.format(google_cloud_pipeline_components.__version__))\""
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "d0a15440"
+      },
+      "outputs": [],
+      "source": [
+        "import os\n",
+        "\n",
+        "if not os.getenv(\"IS_TESTING\"):\n",
+        "    # Automatically restart kernel after installs\n",
+        "    import IPython\n",
+        "\n",
+        "    app = IPython.Application.instance()\n",
+        "    app.kernel.do_shutdown(True)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "B9IYalYObAbY"
+      },
+      "source": [
+        "## Before you begin\n",
+        "\n",
+        "### Set up your Google Cloud project\n",
+        "\n",
+        "**The following steps are required, regardless of your notebook environment.**\n",
+        "\n",
+        "1. [Select or create a Google Cloud project](https://console.cloud.google.com/cloud-resource-manager). When you first create an account, you get a $300 free credit towards your compute/storage costs.\n",
+        "\n",
+        "2. [Make sure that billing is enabled for your project](https://cloud.google.com/billing/docs/how-to/modify-project).\n",
+        "\n",
+        "3. [Enable the Vertex AI API](https://console.cloud.google.com/flows/enableapi?apiid=aiplatform.googleapis.com,storage.googleapis.com).\n",
+        "\n",
+        "4. If you are running this notebook locally, you need to install the [Cloud SDK](https://cloud.google.com/sdk)."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "VA_kzAIIj2G_"
+      },
+      "source": [
+        "### Authenticate your Google Cloud account\n",
+        "\n",
+        "**If you are using Vertex AI Workbench Notebooks**, your environment is already\n",
+        "authenticated. Skip this step.\n",
+        "\n",
+        "**If you are using Colab**, run the cell below and follow the instructions\n",
+        "when prompted to authenticate your account via oAuth.\n",
+        "\n",
+        "**Otherwise**, follow these steps:\n",
+        "\n",
+        "1. In the Cloud Console, go to the [**Create service account key**\n",
+        "   page](https://console.cloud.google.com/apis/credentials/serviceaccountkey).\n",
+        "\n",
+        "2. Click **Create service account**.\n",
+        "\n",
+        "3. In the **Service account name** field, enter a name, and\n",
+        "   click **Create**.\n",
+        "\n",
+        "4. In the **Grant this service account access to project** section, click the **Role** drop-down list. Type \"Vertex AI\"\n",
+        "into the filter box, and select\n",
+        "   **Vertex AI Administrator**. Type \"Storage Object Admin\" into the filter box, and select **Storage Object Admin**.\n",
+        "\n",
+        "5. Click *Create*. A JSON file that contains your key downloads to your\n",
+        "local environment.\n",
+        "\n",
+        "6. Enter the path to your service account key as the\n",
+        "`GOOGLE_APPLICATION_CREDENTIALS` variable in the cell below and run the cell."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "PyQmSRbKA8r-"
+      },
+      "outputs": [],
+      "source": [
+        "# If you are running this notebook in Colab, run this cell and follow the\n",
+        "# instructions to authenticate your GCP account. This provides access to your\n",
+        "# Cloud Storage bucket and lets you submit training jobs and prediction\n",
+        "# requests.\n",
+        "\n",
+        "import os\n",
+        "import sys\n",
+        "\n",
+        "# If on Vertex AI Workbench, then don't execute this code\n",
+        "IS_COLAB = \"google.colab\" in sys.modules\n",
+        "if not os.path.exists(\"/opt/deeplearning/metadata/env_version\") and not os.getenv(\n",
+        "    \"DL_ANACONDA_HOME\"\n",
+        "):\n",
+        "    if \"google.colab\" in sys.modules:\n",
+        "        from google.colab import auth as google_auth\n",
+        "\n",
+        "        google_auth.authenticate_user()\n",
+        "\n",
+        "    # If you are running this notebook locally, replace the string below with the\n",
+        "    # path to your service account key and run this cell to authenticate your GCP\n",
+        "    # account.\n",
+        "    elif not os.getenv(\"IS_TESTING\"):\n",
+        "        %env GOOGLE_APPLICATION_CREDENTIALS ''"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "set_service_account"
+      },
+      "source": [
+        "### Set project ID\n",
+        "\n",
+        "Set your project ID here. If you don't know this, the following snippet attempts to deterine this from your gcloud config. Please continue only if the notebook can see your desired project."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "AkqEd5Gin9mn"
+      },
+      "outputs": [],
+      "source": [
+        "PROJECT_ID = \"your-project-id\"  # @param {type:\"string\"}\n",
+        "if PROJECT_ID == \"\" or PROJECT_ID is None or PROJECT_ID == \"[your-project-id]\":\n",
+        "    # Get your GCP project id from gcloud\n",
+        "    shell_output = !gcloud config list --format 'value(core.project)' 2>/dev/null\n",
+        "    PROJECT_ID = shell_output[0]\n",
+        "print(\"Project ID:\", PROJECT_ID)"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "OVO_gUqpFEP2"
+      },
+      "outputs": [],
+      "source": [
+        "!gcloud config set project $PROJECT_ID"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "a27d4cee"
+      },
+      "source": [
+        "### Setup project information\n",
+        "\n",
+        "Enter information about your project and datasets here."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "7e9477a2"
+      },
+      "outputs": [],
+      "source": [
+        "REGION = \"us\"  # @param {type:\"string\"}\n",
+        "LOCATION = \"us-central1\"  # @param {type:\"string\"}\n",
+        "TRAINING_DATA_LOCATION = \"gs://your-training-data-location\"  # @param {type:\"string\"}\n",
+        "TASK_TYPE = \"CLASSIFICATION\"  # @param [\"CLASSIFICATION\", \"MULTILABEL_CLASSIFICATION\"]"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "o-MZnHsimbOH"
+      },
+      "outputs": [],
+      "source": [
+        "# Since we are training a custom model, we need to specify the list of possible\n",
+        "# classes/labels.\n",
+        "# e.g, [\"FirstClass\", \"SecondClass\"]\n",
+        "# An additional class \"[UNK]\" will be added to the list indicating that none of\n",
+        "# the specified labels are a match.\n",
+        "CLASS_NAMES = [\"\"]\n",
+        "\n",
+        "# This is a list of GCS URIs; e.g., [\"gs://your-bucket-name-here/your-input-file.jsonl\"].\n",
+        "TEST_DATA_URIS = [\"gs://your-bucket-name-here/your-input-file.jsonl\"]"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "timestamp"
+      },
+      "source": [
+        "#### UUID\n",
+        "\n",
+        "To avoid name collisions with other resources in your project, you can create a UUID with the code below and append it onto the name of the bucket(s) created in this notebook."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "wh9sgzemwLXE"
+      },
+      "outputs": [],
+      "source": [
+        "import random\n",
+        "import string\n",
+        "\n",
+        "\n",
+        "# Generate a uuid of a specifed length(default=8)\n",
+        "def generate_uuid(length: int = 8) -> str:\n",
+        "    return \"\".join(random.choices(string.ascii_lowercase + string.digits, k=length))\n",
+        "\n",
+        "\n",
+        "UUID = generate_uuid()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "bucket:mbsdk"
+      },
+      "source": [
+        "### Create a Cloud Storage bucket\n",
+        "\n",
+        "**The following steps are required, regardless of your notebook environment.**\n",
+        "\n",
+        "When you initialize the Vertex AI SDK for Python, you specify a Cloud Storage staging bucket. The staging bucket is where all the data associated with your dataset and model resources are retained across sessions.\n",
+        "\n",
+        "Set the name of your Cloud Storage bucket below. Bucket names must be globally unique across all Google Cloud projects, including those outside of your organization."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "bucket"
+      },
+      "outputs": [],
+      "source": [
+        "BUCKET_NAME = \"[your-bucket-name]\"  # @param {type:\"string\"}\n",
+        "BUCKET_URI = f\"gs://{BUCKET_NAME}\""
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "autoset_bucket"
+      },
+      "outputs": [],
+      "source": [
+        "if BUCKET_NAME == \"\" or BUCKET_NAME is None or BUCKET_NAME == \"[your-bucket-name]\":\n",
+        "    BUCKET_NAME = PROJECT_ID + \"aip-\" + UUID\n",
+        "    BUCKET_URI = \"gs://\" + BUCKET_NAME"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "create_bucket"
+      },
+      "source": [
+        "**Only if your bucket doesn't already exist**: Run the following cell to create your Cloud Storage bucket."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "dO0NV93IwLXF"
+      },
+      "outputs": [],
+      "source": [
+        "!gsutil mb -l $REGION $BUCKET_URI"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "validate_bucket"
+      },
+      "source": [
+        "Finally, validate access to your Cloud Storage bucket by examining its contents:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "Hg5f2oKBwLXG"
+      },
+      "outputs": [],
+      "source": [
+        "!gsutil ls -al $BUCKET_URI"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "EuFETRptyKXc"
+      },
+      "outputs": [],
+      "source": [
+        "from google.cloud import aiplatform\n",
+        "\n",
+        "aiplatform.init(project=PROJECT_ID, staging_bucket=BUCKET_URI)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "f3a09765"
+      },
+      "source": [
+        "## Create training pipeline"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "89bb4a50"
+      },
+      "source": [
+        "### Import libraries"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "0f361e65"
+      },
+      "outputs": [],
+      "source": [
+        "from google_cloud_pipeline_components.aiplatform import ModelBatchPredictOp\n",
+        "from google_cloud_pipeline_components.experimental import natural_language\n",
+        "from google_cloud_pipeline_components.experimental.evaluation import (\n",
+        "    GetVertexModelOp, ModelEvaluationClassificationOp,\n",
+        "    TargetFieldDataRemoverOp)\n",
+        "from kfp import components\n",
+        "from kfp.v2 import compiler, dsl"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d33c87e4-2ada-4b87-bf75-064247f3162d"
+      },
+      "source": [
+        "### Define constants"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "36ceb9f8"
+      },
+      "outputs": [],
+      "source": [
+        "# Worker pool specs\n",
+        "TRAINING_MACHINE_TYPE = \"n1-highmem-8\"\n",
+        "ACCELERATOR_TYPE = \"NVIDIA_TESLA_T4\"\n",
+        "ACCELERATOR_COUNT = 1\n",
+        "EVAL_MACHINE_TYPE = \"n1-highmem-8\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "zAaMJKrhAe5L"
+      },
+      "source": [
+        "## Define components\n",
+        "\n",
+        "This pipeline is composed from the following components:\n",
+        "\n",
+        "- **train-tfhub-model** - Trains a new Tensorflow model using TFHub layers from pre-built Docker image\n",
+        "- **upload-tensorflow-model-to-google-cloud-vertex-ai** - Uploads resulting model to Vertex AI model registry\n",
+        "- **get-vertex-model** - Gets model that has just been uploaded as an artifact in pipeline\n",
+        "- **convert-dataset-export-for-batch-predict** - Preprocessing component that takes the test dataset exported from Vertex datasets and converts it to a simpler compatible one that is readable from the batch predict component\n",
+        "- **target-field-data-remover** - Removes the target field (i.e., label) in the test dataset for the downstream batch predict component\n",
+        "- **model-batch-predict** - Performs a batch prediction job\n",
+        "- **model-evaluation-classification** - Calculates the evaluation metrics from the above batch predict job and exports the metrics artifact\n"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "DKe2iQNKgpKG"
+      },
+      "outputs": [],
+      "source": [
+        "# Load upload TF model component\n",
+        "upload_tensorflow_model_to_vertex_op = components.load_component_from_url(\n",
+        "    \"https://raw.githubusercontent.com/Ark-kun/pipeline_components/c6a8b67d1ada2cc17665c99ff6b410df588bee28/components/google-cloud/Vertex_AI/Models/Upload_Tensorflow_model/workaround_for_buggy_KFPv2_compiler/component.yaml\"\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "TEnh9Pcx6Xfi"
+      },
+      "source": [
+        "### Define the pipeline\n",
+        "\n",
+        "The pipeline performs the following steps:\n",
+        "- Trains new text classification model\n",
+        "- Uploads model to Vertex AI Model Registry\n",
+        "- Performs preprocessing steps on test dataset export: formats data for batch predcition, removes target field\n",
+        "- Performs batch prediction on preprocessed test data\n",
+        "- Evaluates performance of model based on batch prediction output"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "2a67cde8"
+      },
+      "outputs": [],
+      "source": [
+        "@dsl.pipeline(name=\"text-classification-model\")\n",
+        "def pipeline():\n",
+        "    train_task = natural_language.TrainTextClassificationOp()(\n",
+        "        project=PROJECT_ID,\n",
+        "        location=LOCATION,\n",
+        "        machine_type=TRAINING_MACHINE_TYPE,\n",
+        "        accelerator_type=ACCELERATOR_TYPE,\n",
+        "        accelerator_count=ACCELERATOR_COUNT,\n",
+        "        input_data_path=TRAINING_DATA_LOCATION,\n",
+        "        input_format=\"jsonl\",\n",
+        "        natural_language_task_type=TASK_TYPE,\n",
+        "    )\n",
+        "\n",
+        "    upload_task = upload_tensorflow_model_to_vertex_op(\n",
+        "        model=train_task.outputs[\"model_output\"]\n",
+        "    )\n",
+        "\n",
+        "    get_model_task = GetVertexModelOp(\n",
+        "        model_resource_name=upload_task.outputs[\"model_name\"]\n",
+        "    )\n",
+        "\n",
+        "    classification_type = (\n",
+        "        \"multilabel\" if TASK_TYPE == \"MULTILABEL_CLASSIFICATION\" else \"multiclass\"\n",
+        "    )\n",
+        "\n",
+        "    convert_dataset_task = natural_language.ConvertDatasetExportForBatchPredictOp(\n",
+        "        file_paths=TEST_DATA_URIS, classification_type=classification_type\n",
+        "    )\n",
+        "\n",
+        "    target_field_remover_task = TargetFieldDataRemoverOp(\n",
+        "        project=PROJECT_ID,\n",
+        "        location=LOCATION,\n",
+        "        root_dir=BUCKET_URI,\n",
+        "        gcs_source_uris=convert_dataset_task.outputs[\"output_files\"],\n",
+        "        target_field_name=\"labels\",\n",
+        "        instances_format=\"jsonl\",\n",
+        "    )\n",
+        "\n",
+        "    # Note: ModelBatchPredictOp doesn't support accelerators currently.\n",
+        "    batch_predict_task = ModelBatchPredictOp(\n",
+        "        project=PROJECT_ID,\n",
+        "        location=LOCATION,\n",
+        "        model=get_model_task.outputs[\"model\"],\n",
+        "        job_display_name=\"nl-batch-predict-evaluation\",\n",
+        "        gcs_source_uris=target_field_remover_task.outputs[\"gcs_output_directory\"],\n",
+        "        instances_format=\"jsonl\",\n",
+        "        predictions_format=\"jsonl\",\n",
+        "        gcs_destination_output_uri_prefix=BUCKET_URI,\n",
+        "        machine_type=EVAL_MACHINE_TYPE,\n",
+        "    )\n",
+        "\n",
+        "    # Note: Because we're running a custom training pipeline, the model source\n",
+        "    # is detected as Custom and thus it doesn't use AutoML NL's default settings\n",
+        "    # and fails if class_labels is excluded.\n",
+        "    ModelEvaluationClassificationOp(\n",
+        "        project=PROJECT_ID,\n",
+        "        location=LOCATION,\n",
+        "        root_dir=BUCKET_URI,\n",
+        "        class_labels=CLASS_NAMES + [\"[UNK]\"],\n",
+        "        predictions_gcs_source=batch_predict_task.outputs[\"gcs_output_directory\"],\n",
+        "        predictions_format=\"jsonl\",\n",
+        "        prediction_label_column=\"prediction.displayNames\",\n",
+        "        prediction_score_column=\"prediction.confidences\",\n",
+        "        ground_truth_gcs_source=convert_dataset_task.outputs[\"output_files\"],\n",
+        "        ground_truth_format=\"jsonl\",\n",
+        "        target_field_name=\"labels\",\n",
+        "        classification_type=TASK_TYPE,\n",
+        "    )"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "3211ba19"
+      },
+      "source": [
+        "### Compile the pipeline"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "c368c73f"
+      },
+      "outputs": [],
+      "source": [
+        "compiler.Compiler().compile(pipeline, \"nl_pipeline.json\")"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "l_Vxwz5cdF5f"
+      },
+      "source": [
+        "Running the above line will generate a file locally or in Colab's directory."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "ax0jOxIaholy"
+      },
+      "source": [
+        "### Run the pipeline\n",
+        "\n",
+        "This sends a create pipeline job request to Vertex Pipelines. Note that this  task run synchronously and may take a while to complete.\n",
+        "\n",
+        "You may view the progress of the job at any time by clicking on the generated links (after \"View Pipeline Job\" in the console output of the cell below). Once the pipeline finishes, you may examine the artifacts produced from this pipeline."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "Wfs7QOSxhp_n"
+      },
+      "outputs": [],
+      "source": [
+        "job = aiplatform.PipelineJob(\n",
+        "    display_name=\"nl_pipeline\",\n",
+        "    template_path=\"nl_pipeline.json\",\n",
+        "    location=LOCATION,\n",
+        "    enable_caching=True,\n",
+        "    parameter_values={},\n",
+        ")\n",
+        "\n",
+        "job.run()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "UIyGPaihWJWn"
+      },
+      "source": [
+        "Once the pipeline successfully finishes, go to the pipeline and examine the resulting metrics artifacts for the results. Otherwise, refer to the failing step(s) in the pipeline to determine the cause of any errors."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "OoexTJTy9jnH"
+      },
+      "source": [
+        "## View model evaluation results\n",
+        "\n",
+        "To check the results of evaluation after pipeline execution, find the \"model-evaluation-classification\" subdirectory in the Cloud Storage bucket created by this pipeline. You may also run the following to directly output the contents of the metrics file:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "h9EqPCQF9lN9"
+      },
+      "outputs": [],
+      "source": [
+        "import tensorflow as tf\n",
+        "\n",
+        "EVAL_TASK_NAME = \"model-evaluation-classification\"\n",
+        "PROJECT_NUMBER = job.gca_resource.name.split(\"/\")[1]\n",
+        "for _ in range(len(job.gca_resource.job_detail.task_details)):\n",
+        "    TASK_ID = job.gca_resource.job_detail.task_details[_].task_id\n",
+        "    EVAL_METRICS = (\n",
+        "        BUCKET_URI\n",
+        "        + \"/\"\n",
+        "        + PROJECT_NUMBER\n",
+        "        + \"/\"\n",
+        "        + job.name\n",
+        "        + \"/\"\n",
+        "        + EVAL_TASK_NAME\n",
+        "        + \"_\"\n",
+        "        + str(TASK_ID)\n",
+        "        + \"/executor_output.json\"\n",
+        "    )\n",
+        "    if tf.io.gfile.exists(EVAL_METRICS):\n",
+        "        ! gsutil cat $EVAL_METRICS"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "TpV-iwP9qw9c"
+      },
+      "source": [
+        "## Cleaning up\n",
+        "\n",
+        "To clean up the resources used by this pipeline, run the command below:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "sx_vKniMq9ZX"
+      },
+      "outputs": [],
+      "source": [
+        "# Delete GCS bucket.\n",
+        "!gsutil -m rm -r {BUCKET_URI}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "UMuyzrnZLoUa"
+      },
+      "source": [
+        "# Next steps\n",
+        "\n",
+        "For an alternate approach, please check out the [\"ready-to-go\" text classification pipeline](https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/master/notebooks/community/pipelines/google_cloud_pipeline_components_ready_to_go_text_classification_pipeline.ipynb). This pipeline exposes the model logic for further customization if needed, and adds an additional pipeline step to deploy the model to enable online predictions."
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "collapsed_sections": [
+        "d975e698c9a4",
+        "08d289fa873f",
+        "d33c87e4-2ada-4b87-bf75-064247f3162d",
+        "3211ba19",
+        "TpV-iwP9qw9c",
+        "UMuyzrnZLoUa"
+      ],
+      "name": "google_cloud_pipeline_components_cloud_natural_language_pipeline.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/community/sdk/sdk_pytorch_torchrun_custom_container_training_imagenet.ipynb

@@ -567,7 +567,7 @@
       "outputs": [],
       "source": [
         "%%writefile trainer/Dockerfile\n",
-        "FROM gcr.io/deeplearning-platform-release/pytorch-gpu.1-12\n",
+        "FROM gcr.io/deeplearning-platform-release/pytorch-gpu.1-13:m102\n",
         "\n",
         "RUN curl https://packages.cloud.google.com/apt/doc/apt-key.gpg | apt-key add - && \\\n",
         "    # Install reduction server plugin on GPU containers. google-fast-socket is\n",
@@ -588,16 +588,31 @@
         "RUN apt-get update -y && \\\n",
         "    apt-get install -y curl gnupg telnet nano net-tools iputils-ping\n",
         "\n",
+        "# Set ETCD version\n",
+        "ARG ETCD_VER=v2.3.0\n",
+        "# Choose either URL\n",
+        "ARG GOOGLE_URL=https://storage.googleapis.com/etcd\n",
+        "ARG GITHUB_URL=https://github.com/etcd-io/etcd/releases/download\n",
+        "# Set ETCD URL to download from\n",
+        "ARG DOWNLOAD_URL=$GOOGLE_URL\n",
         "\n",
+        "# Install ETCD\n",
+        "RUN mkdir -p /tmp/etcd-download-test && \\\n",
+        "    curl -L ${DOWNLOAD_URL}/${ETCD_VER}/etcd-${ETCD_VER}-linux-amd64.tar.gz -o /tmp/etcd-${ETCD_VER}-linux-amd64.tar.gz && \\\n",
+        "    tar xzvf /tmp/etcd-${ETCD_VER}-linux-amd64.tar.gz -C /tmp/etcd-download-test --strip-components=1 && \\\n",
+        "    rm -f /tmp/etcd-${ETCD_VER}-linux-amd64.tar.gz\n",
+        "\n",
+        "# Copy training application code\n",
         "COPY . /trainer\n",
         "\n",
         "WORKDIR /trainer\n",
         "\n",
+        "# Install dependencies\n",
         "RUN pip install -r requirements.txt\n",
         "\n",
         "RUN chmod 777 main.sh\n",
         "\n",
-        "# download data to the container\n",
+        "# Download data to the container\n",
         "RUN wget -q -P /trainer/data https://image-net.org/data/tiny-imagenet-200.zip\n",
         "RUN unzip -q /trainer/data/tiny-imagenet-200.zip\n",
         "RUN rm /trainer/data/tiny-imagenet-200.zip\n",
@@ -614,8 +629,8 @@
       "outputs": [],
       "source": [
         "%%writefile trainer/requirements.txt\n",
-        "torch==1.12.0\n",
-        "torchvision==0.13.0\n",
+        "torch==1.13.0\n",
+        "torchvision==0.14.0\n",
         "tensorboard==2.5.0\n",
         "protobuf==3.20.*\n",
         "python-etcd\n",
@@ -675,30 +690,17 @@
         "setup_etcd() {\n",
         "    HOST_IP=$1\n",
         "    # Start a local instane of ETCD v2 \n",
-        "    ETCD_VER=v2.3.0 #v3.5.6\n",
         "    export ETCD_ENABLE_V2=true\n",
         "    export ETCDCTL_API=2\n",
         "\n",
-        "    # choose either URL\n",
-        "    GOOGLE_URL=https://storage.googleapis.com/etcd\n",
-        "    GITHUB_URL=https://github.com/etcd-io/etcd/releases/download\n",
-        "    DOWNLOAD_URL=${GOOGLE_URL}\n",
-        "\n",
-        "    rm -f /tmp/etcd-${ETCD_VER}-linux-amd64.tar.gz\n",
-        "    rm -rf /tmp/etcd-download-test && mkdir -p /tmp/etcd-download-test\n",
-        "\n",
-        "    curl -L ${DOWNLOAD_URL}/${ETCD_VER}/etcd-${ETCD_VER}-linux-amd64.tar.gz -o /tmp/etcd-${ETCD_VER}-linux-amd64.tar.gz\n",
-        "    tar xzvf /tmp/etcd-${ETCD_VER}-linux-amd64.tar.gz -C /tmp/etcd-download-test --strip-components=1\n",
-        "    rm -f /tmp/etcd-${ETCD_VER}-linux-amd64.tar.gz\n",
-        "\n",
         "    /tmp/etcd-download-test/etcd --name s1 --data-dir /tmp/etcd-download-test/s1  \\\n",
         "    --listen-client-urls http://0.0.0.0:2379 --advertise-client-urls http://$HOST_IP:2379 \\\n",
         "    --listen-peer-urls http://0.0.0.0:2380 --initial-advertise-peer-urls http://$HOST_IP:2380 \\\n",
         "    --initial-cluster s1=http://$HOST_IP:2380 --initial-cluster-token tkn \\\n",
         "    --initial-cluster-state new &> /tmp/etcd-download-test/node.log &\n",
         "\n",
-        "    sudo /tmp/etcd-download-test/etcd --version\n",
-        "    sudo /tmp/etcd-download-test/etcdctl --version\n",
+        "    /tmp/etcd-download-test/etcd --version\n",
+        "    /tmp/etcd-download-test/etcdctl --version\n",
         "}\n",
         "\n",
         "\n",

notebooks/community/vertex_endpoints/optimized_tensorflow_runtime/benchmark.py

@@ -112,7 +112,7 @@ def benchmark(
 
     results = []
     for qps in qps_list:
-        num_requests = max(qps * duration_sec, 10)
+        num_requests = int(max(qps * duration_sec, 10))
         requests_for_qps = list(
             itertools.islice(itertools.cycle(requests), num_requests)
         )

notebooks/community/vertex_endpoints/torchserve/dreambooth_stablediffusion.ipynb

@@ -0,0 +1,987 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "18ebbd838e32"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2023 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "# https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "219f1b1fe8fe"
+      },
+      "source": [
+        "# Deploy and host a Stable Diffusion model on Vertex AI\n"
+      ]
+    },
+    {
+      "attachments": {},
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "JAPoU8Sm5E6e"
+      },
+      "source": [
+        "<table align=\"left\">\n",
+        "\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/vertex_endpoints/torchserve/dreambooth_stablediffusion.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/vertex_endpoints/torchserve/dreambooth_stablediffusion.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        " <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/workbench/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/vertex_endpoints/torchserve/dreambooth_stablediffusion.ipynb\">\n",
+        "      <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "      Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "fce05a8186d6"
+      },
+      "source": [
+        "## Overview\n",
+        "\n",
+        "This notebook demonstrates how to deploy and host a fine-tuned [Stable Diffusion 1.5](https://huggingface.co/runwayml/stable-diffusion-v1-5) model on Vertex AI. For hosting, you use the PyTorch 3 container built for Vertex AI with [TorchServe](https://pytorch.org/serve/index.html)."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "c76216b03fec"
+      },
+      "source": [
+        "### Objective\n",
+        "\n",
+        "In this tutorial, you learn how to host and deploy a Stable Diffusion 1.5 model on Vertex AI.\n",
+        "\n",
+        "This tutorial uses the following Google Cloud ML services:\n",
+        "\n",
+        "+ Vertex AI `Model` resource\n",
+        "+ Vertex AI `Endpoint` resource\n",
+        "\n",
+        "The steps performed include:\n",
+        "\n",
+        "+ Create a `torchserve` handler for responding to prediction requests.\n",
+        "+ Upload a Stable Diffusion 1.5 model on a prebuilt PyTorch container in Vertex AI.\n",
+        "+ Deploy a model to a Vertex AI Endpoint.\n",
+        "+ Send requests to the endpoint and parse the responses using Vertex AI Prediction service."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "c6deba5a8557"
+      },
+      "source": [
+        "### Model\n",
+        "\n",
+        "This notebook uses a collection of model artifacts fine-tuned to generate images of a small dog. These are the same images used in the original [DreamBooth paper](https://dreambooth.github.io/)."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "911dc651ea9c"
+      },
+      "source": [
+        "### Costs\n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI models\n",
+        "* Vertex AI endpoints\n",
+        "* Vertex AI prediction\n",
+        "* Cloud Storage\n",
+        "* (Optionally) Vertex AI Workbench\n",
+        "\n",
+        "Learn about [Vertex AI pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage pricing](https://cloud.google.com/storage/pricing), and use the [Pricing Calculator](https://cloud.google.com/products/calculator/) to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0d36bd3d53fa"
+      },
+      "source": [
+        "## Hardware requirements\n",
+        "\n",
+        "This notebook requires that you use a GPU with a sufficient amount of VRAM available. It was tested on a `NVIDIA Tesla A100 GPU` with 85 GB of VRAM. Run the following cell to ensure that you have the correct hardware configuration."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "3dd4022552e5"
+      },
+      "outputs": [],
+      "source": [
+        "!nvidia-smi --query-gpu=name,memory.total,memory.free --format=\"csv,noheader\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "a782627e5f73"
+      },
+      "source": [
+        "### Create a user-managed notebook on Vertex AI\n",
+        "\n",
+        "If you are using Vertex AI Workbench, you can create a notebook with the correct configuration by doing the following:\n",
+        "\n",
+        "+ Go to [Vertex AI Workbench](https://console.cloud.google.com/vertex-ai/workbench/user-managed) in the Google Cloud Console.\n",
+        "+ Click **New Notebook** and then click **PyTorch 1.13** > **With 1 NVIDIA T4**.\n",
+        "+ In the **New notebook** dialog box, click **Advanced Options**. The **Create a user-managed notebook** page opens up.\n",
+        "+ In the **Create a user-managed notebook** page, do the following:\n",
+        "  * In the **Notebook name** box, type a name for your notebook, for example \"my-stablediffusion-nb\".\n",
+        "  * In the **Machine type** drop-down, select **A2 highgpu** > **a2-highgpu-1g**.\n",
+        "  * In the **GPU type** drop-down, select **NVIDIA Tesla A100**.\n",
+        "  * Check the box next to **Install NVIDIA GPU driver automatically for me**\n",
+        "  * Expand **Disk(s)** and do the following:\n",
+        "    - Under **Boot disk type**, select **SSD Persistent Disk**.\n",
+        "    - Under **Data disk type**, select **SSD Persistent Disk**.\n",
+        "  * Click **Create**."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "FAPoU8Sm5E6e"
+      },
+      "source": [
+        "<div style=\"background:#feefe3; padding:5px; color:#aa0000\">\n",
+        "<strong>Caution:</strong> Using a Vertex AI Workbench notebook with the above configuration can increase your costs significantly. You can estimate your costs using the <a href=\"https://cloud.google.com/products/calculator\"><u>costs calculator</u></a>.</div>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0bb4201cc99a"
+      },
+      "source": [
+        "## Installation\n",
+        "\n",
+        "Install the following packages required to execute this notebook.\n",
+        "\n",
+        "**Note**: You might need to change the version of PyTorch (`torch`) installed by `pip`."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "9c769df171a6"
+      },
+      "outputs": [],
+      "source": [
+        "%%writefile requirements.txt\n",
+        "diffusers\n",
+        "ftfy\n",
+        "google-cloud-aiplatform\n",
+        "gradio\n",
+        "ninja\n",
+        "tensorboard==1.15.0\n",
+        "torch\n",
+        "torchaudio\n",
+        "torchvision\n",
+        "torchserve\n",
+        "torch-model-archiver\n",
+        "torch-workflow-archiver\n",
+        "transformers"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "e46804ac90d8"
+      },
+      "outputs": [],
+      "source": [
+        "%pip install -r requirements.txt"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "58707a750154"
+      },
+      "source": [
+        "### Colab only: Uncomment the following cell to restart the kernel."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "77c11549298a"
+      },
+      "outputs": [],
+      "source": [
+        "# Automatically restart kernel after installs so that your environment can access the new packages\n",
+        "# import IPython\n",
+        "\n",
+        "# app = IPython.Application.instance()\n",
+        "# app.kernel.do_shutdown(True)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "294df346a918"
+      },
+      "source": [
+        "## Before you begin"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "BF1j6f9HApxa"
+      },
+      "source": [
+        "### Set up your Google Cloud project\n",
+        "\n",
+        "**The following steps are required, regardless of your notebook environment.**\n",
+        "\n",
+        "1. [Select or create a Google Cloud project](https://console.cloud.google.com/cloud-resource-manager). When you first create an account, you get a $300 free credit towards your compute/storage costs.\n",
+        "\n",
+        "2. [Make sure that billing is enabled for your project](https://cloud.google.com/billing/docs/how-to/modify-project).\n",
+        "\n",
+        "3. [Enable the Vertex AI API](https://console.cloud.google.com/flows/enableapi?apiid=aiplatform.googleapis.com).\n",
+        "\n",
+        "4. If you are running this notebook locally, you need to install the [Cloud SDK](https://cloud.google.com/sdk)."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "WReHDGG5g0XY"
+      },
+      "source": [
+        "#### Set your project ID\n",
+        "\n",
+        "**If you don't know your project ID**, try the following:\n",
+        "* Run `gcloud config list`.\n",
+        "* Run `gcloud projects list`.\n",
+        "* See the support page: [Locate the project ID](https://support.google.com/googleapi/answer/7014113)"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "oM1iC_MfAts1"
+      },
+      "outputs": [],
+      "source": [
+        "PROJECT_ID = \"[your-project-id]\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# Set the project id\n",
+        "! gcloud config set project {PROJECT_ID}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "region"
+      },
+      "source": [
+        "#### Region\n",
+        "\n",
+        "You can also change the `REGION` variable used by Vertex AI. Learn more about [Vertex AI regions](https://cloud.google.com/vertex-ai/docs/general/locations)."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "region"
+      },
+      "outputs": [],
+      "source": [
+        "REGION = \"us-central1\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "sBCra4QMA2wR"
+      },
+      "source": [
+        "### Authenticate your Google Cloud account\n",
+        "\n",
+        "Depending on your Jupyter environment, you may have to manually authenticate. Follow the relevant instructions below."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "74ccc9e52986"
+      },
+      "source": [
+        "**1. Vertex AI Workbench**\n",
+        "* Do nothing as you are already authenticated."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "de775a3773ba"
+      },
+      "source": [
+        "**2. Local JupyterLab instance, uncomment and run:**"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "254614fa0c46"
+      },
+      "outputs": [],
+      "source": [
+        "# ! gcloud auth login"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "ef21552ccea8"
+      },
+      "source": [
+        "**3. Colab, uncomment and run:**"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "603adbbf0532"
+      },
+      "outputs": [],
+      "source": [
+        "# from google.colab import auth\n",
+        "# auth.authenticate_user()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "f6b2ccc891ed"
+      },
+      "source": [
+        "**4. Service account or other**\n",
+        "* See how to grant Cloud Storage permissions to your service account at https://cloud.google.com/storage/docs/gsutil/commands/iam#ch-examples."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "cb5c4ca3e851"
+      },
+      "source": [
+        "### Import libraries"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "7348591eda51"
+      },
+      "outputs": [],
+      "source": [
+        "import base64\n",
+        "import math\n",
+        "\n",
+        "import torch\n",
+        "from diffusers import StableDiffusionPipeline\n",
+        "from google.cloud import aiplatform\n",
+        "from IPython import display\n",
+        "from PIL import Image\n",
+        "from torch import autocast"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "697566b5f660"
+      },
+      "source": [
+        "## Optional: View model inferences\n",
+        "\n",
+        "Before uploading the model to Vertex AI, you can review the expected output from the model. The model used in this notebook is available for your use and can be downloaded from Cloud Storage. This download may take a few minutes to complete."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "d63df8d91215"
+      },
+      "outputs": [],
+      "source": [
+        "!gsutil -m cp gs://cloud-samples-data/vertex-ai/model-deployment/models/stable-diffusion/model_artifacts.zip \\\n",
+        "   ."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "89d613ae9573"
+      },
+      "outputs": [],
+      "source": [
+        "!unzip model_artifacts.zip"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "9192ea4f3b57"
+      },
+      "source": [
+        "### Create new images\n",
+        "\n",
+        "With everything in place, you can now generate new images from the Stable Diffusion model. First you must load your model into a `StableDiffusionPipeline`."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "cd1f30223b79"
+      },
+      "outputs": [],
+      "source": [
+        "model_path = \"model_artifacts\"\n",
+        "\n",
+        "pipe = StableDiffusionPipeline.from_pretrained(\n",
+        "    model_path, torch_dtype=torch.float16\n",
+        ").to(\"cuda\")\n",
+        "\n",
+        "g_cuda = None"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "dd91520f58cf"
+      },
+      "outputs": [],
+      "source": [
+        "g_cuda = torch.Generator(device=\"cuda\")\n",
+        "seed = 52362\n",
+        "g_cuda.manual_seed(seed)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "1e9c5d734b9f"
+      },
+      "source": [
+        "With the model loaded into a `StableDiffusionPipeline`, you can now generate results (inferences) from the model. Each set of inference requires an input (called a [prompt](https://learnprompting.org/)) that specifies what the model should create.\n",
+        "\n",
+        "You can also vary other inputs into the model, as shown in the following cell."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "a28b73de55ce"
+      },
+      "outputs": [],
+      "source": [
+        "prompt = \"photo of examplePup dog in a Monet style\"\n",
+        "\n",
+        "num_samples = 4\n",
+        "num_batches = 1\n",
+        "num_columns = 2\n",
+        "guidance_scale = 10\n",
+        "num_inference_steps = 50\n",
+        "height = 512\n",
+        "width = 512"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "11a7bb79de59"
+      },
+      "outputs": [],
+      "source": [
+        "def image_grid(imgs, cols):\n",
+        "    total = len(imgs)\n",
+        "    rows = math.ceil(total / cols)\n",
+        "\n",
+        "    w, h = imgs[0].size\n",
+        "    grid = Image.new(\"RGB\", size=(cols * w, rows * h))\n",
+        "    grid_w, grid_h = grid.size\n",
+        "\n",
+        "    for i, img in enumerate(imgs):\n",
+        "        grid.paste(img, box=(i % cols * w, i // cols * h))\n",
+        "    return grid\n",
+        "\n",
+        "\n",
+        "all_images = []\n",
+        "for _ in range(num_batches):\n",
+        "    with autocast(\"cuda\"):\n",
+        "        images = pipe(\n",
+        "            [prompt] * num_samples,\n",
+        "            height=height,\n",
+        "            width=width,\n",
+        "            num_inference_steps=num_inference_steps,\n",
+        "            guidance_scale=guidance_scale,\n",
+        "        ).images\n",
+        "        all_images.extend(images)\n",
+        "\n",
+        "\n",
+        "grid = image_grid(all_images, num_columns)\n",
+        "grid"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "bbc72963ff88"
+      },
+      "source": [
+        "## Deploy the model to Vertex AI\n",
+        "\n",
+        "You can host your Stable Diffusion 1.5 model on a Vertex AI endpoint where you can get inferences from it online. Uploading your model is a four step process: \n",
+        "\n",
+        "1. Create a custom TorchServe handler.\n",
+        "1. Upload the model artifacts onto Cloud Storage.\n",
+        "2. Create a Vertex AI model with the model artifacts and a prebuilt PyTorch container image.\n",
+        "3. Deploy the Vertex AI model onto an endpoint."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "eafbb0e0-40e6-43a0-a38e-edc54323da51"
+      },
+      "source": [
+        "### Create the custom TorchServe handler\n",
+        "\n",
+        "The model deployed to Vertex AI uses [TorchServe](https://pytorch.org/serve/) to handle requests and return responses from the model. You must create a custom TorchServe handler to include in with the model artifacts uploaded to Vertex AI.\n",
+        "\n",
+        "The handler file should be included in the directory with the other model artifacts."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "94567a87-9d74-4c87-a749-306ddaf01b61"
+      },
+      "outputs": [],
+      "source": [
+        "%%writefile model_artifacts/handler.py\n",
+        "\n",
+        "\"\"\"Customized handler for Stable Diffusion 1.5.\"\"\"\n",
+        "import base64\n",
+        "import logging\n",
+        "from io import BytesIO\n",
+        "\n",
+        "import torch\n",
+        "from diffusers import EulerDiscreteScheduler\n",
+        "from diffusers import StableDiffusionPipeline\n",
+        "from ts.torch_handler.base_handler import BaseHandler\n",
+        "\n",
+        "logger = logging.getLogger(__name__)\n",
+        "model_id = 'runwayml/stable-diffusion-v1-5'\n",
+        "\n",
+        "\n",
+        "class ModelHandler(BaseHandler):\n",
+        "\n",
+        "  def __init__(self):\n",
+        "    self.initialized = False\n",
+        "    self.map_location = None\n",
+        "    self.device = None\n",
+        "    self.use_gpu = True\n",
+        "    self.store_avg = True\n",
+        "    self.pipe = None\n",
+        "\n",
+        "  def initialize(self, context):\n",
+        "    \"\"\"Initializes the pipe.\"\"\"\n",
+        "    properties = context.system_properties\n",
+        "    gpu_id = properties.get('gpu_id')\n",
+        "\n",
+        "    self.map_location, self.device, self.use_gpu = \\\n",
+        "      ('cuda', torch.device('cuda:' + str(gpu_id)),\n",
+        "       True) if torch.cuda.is_available() else \\\n",
+        "        ('cpu', torch.device('cpu'), False)\n",
+        "\n",
+        "    # Use the Euler scheduler here instead\n",
+        "    scheduler = EulerDiscreteScheduler.from_pretrained(model_id,\n",
+        "                                                       subfolder='scheduler')\n",
+        "    pipe = StableDiffusionPipeline.from_pretrained(model_id,\n",
+        "                                                   scheduler=scheduler,\n",
+        "                                                   torch_dtype=torch.float16)\n",
+        "    pipe = pipe.to('cuda')\n",
+        "    # Uncomment the following line to reduce the GPU memory usage.\n",
+        "    # pipe.enable_attention_slicing()\n",
+        "    self.pipe = pipe\n",
+        "\n",
+        "    self.initialized = True\n",
+        "\n",
+        "  def preprocess(self, requests):\n",
+        "    \"\"\"Noting to do here.\"\"\"\n",
+        "    logger.info('requests: %s', requests)\n",
+        "    return requests\n",
+        "\n",
+        "  def inference(self, preprocessed_data, *args, **kwargs):\n",
+        "    \"\"\"Run the inference.\"\"\"\n",
+        "    images = []\n",
+        "    for pd in preprocessed_data:\n",
+        "      prompt = pd['prompt']\n",
+        "      images.extend(self.pipe(prompt).images)\n",
+        "    return images\n",
+        "\n",
+        "  def postprocess(self, output_batch):\n",
+        "    \"\"\"Converts the images to base64 string.\"\"\"\n",
+        "    postprocessed_data = []\n",
+        "    for op in output_batch:\n",
+        "      fp = BytesIO()\n",
+        "      op.save(fp, format='JPEG')\n",
+        "      postprocessed_data.append(base64.b64encode(fp.getvalue()).decode('utf-8'))\n",
+        "      fp.close()\n",
+        "    return postprocessed_data\n"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "6ace1dac0af0"
+      },
+      "source": [
+        "After creating the handler file, you must package the handler as a model archiver (MAR) file. The output file must be named 'model.mar'."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "67707f95d440"
+      },
+      "outputs": [],
+      "source": [
+        "!torch-model-archiver \\\n",
+        "  -f \\\n",
+        "  --model-name model \\\n",
+        "  --version 1.0 \\\n",
+        "  --handler model_artifacts/handler.py \\\n",
+        "  --export-path model_artifacts"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "ffab030f4bc8"
+      },
+      "source": [
+        "### Upload the model artifacts to Cloud Storage\n",
+        "\n",
+        "Create a new folder in your Cloud Storage bucket to hold the model artifacts"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "MzGDU7TWdts_"
+      },
+      "outputs": [],
+      "source": [
+        "BUCKET_NAME = \"your-bucket-name-unique\"  # @param {type:\"string\"}\n",
+        "BUCKET_URI = f\"gs://{BUCKET_NAME}/\"\n",
+        "FULL_GCS_PATH = f\"{BUCKET_URI}model_artifacts\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "-EcIXiGsCePi"
+      },
+      "source": [
+        "**Only if your bucket doesn't already exist**: Run the following cell to create your Cloud Storage bucket."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "NIq7R4HZCfIc"
+      },
+      "outputs": [],
+      "source": [
+        "! gsutil mb -l $REGION -p $PROJECT_ID $BUCKET_URI"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "971232e28657"
+      },
+      "source": [
+        "Next, upload the model archive file and your trained Stable Diffusion 1.5 model to the folder on Cloud Storage."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "ef6baf44c808"
+      },
+      "outputs": [],
+      "source": [
+        "!gsutil cp -r model_artifacts $BUCKET_URI"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "402370ca9396"
+      },
+      "source": [
+        "### Create the Vertex AI model\n",
+        "\n",
+        "Once you've uploaded the model artifacts into a Cloud Storage bucket, you can create a new Vertex AI model. This notebook uses the [Vertex AI SDK](https://cloud.google.com/vertex-ai/docs/start/use-vertex-ai-python-sdk) to create the model."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b6c58a74a0fd"
+      },
+      "outputs": [],
+      "source": [
+        "PYTORCH_PREDICTION_IMAGE_URI = (\n",
+        "    \"us-docker.pkg.dev/vertex-ai/prediction/pytorch-gpu.1-12:latest\"\n",
+        ")\n",
+        "APP_NAME = \"my-stable-diffusion\"\n",
+        "VERSION = 1\n",
+        "MODEL_DISPLAY_NAME = \"stable_diffusion_1_5-unique\"\n",
+        "MODEL_DESCRIPTION = \"stable_diffusion_1_5 container\"\n",
+        "ENDPOINT_DISPLAY_NAME = f\"{APP_NAME}-endpoint\""
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "07c3503a1a2e"
+      },
+      "outputs": [],
+      "source": [
+        "aiplatform.init(project=PROJECT_ID, location=REGION, staging_bucket=BUCKET_NAME)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "FAPoU8Sm5E6e"
+      },
+      "source": [
+        "<div style=\"background:#e3effe; padding:5px; color:#0000aa\">\n",
+        "<strong>Note:</strong> The next cell fails if you haven't <a href=\"https://console.cloud.google.com/flows/enableapi?apiid=aiplatform.googleapis.com\"><u>enabled the Vertex API</u></a>.</div>"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "a776324dd16f"
+      },
+      "outputs": [],
+      "source": [
+        "model = aiplatform.Model.upload(\n",
+        "    display_name=MODEL_DISPLAY_NAME,\n",
+        "    description=MODEL_DESCRIPTION,\n",
+        "    serving_container_image_uri=PYTORCH_PREDICTION_IMAGE_URI,\n",
+        "    artifact_uri=FULL_GCS_PATH,\n",
+        ")\n",
+        "\n",
+        "model.wait()\n",
+        "\n",
+        "print(model.display_name)\n",
+        "print(model.resource_name)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0fbc3d371574"
+      },
+      "source": [
+        "### Deploy the model to an endpoint\n",
+        "\n",
+        "To get online preductions from your Stable Diffusion 2.0 model, you must [deploy it to a Vertex AI endpoint](https://cloud.google.com/vertex-ai/docs/predictions/overview). You can again use the Vertex AI SDK to create the endpoint and deploy your model."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "ab29f0a770cb"
+      },
+      "outputs": [],
+      "source": [
+        "endpoint = aiplatform.Endpoint.create(display_name=ENDPOINT_DISPLAY_NAME)"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "25f703df88c7"
+      },
+      "outputs": [],
+      "source": [
+        "model.deploy(\n",
+        "    endpoint=endpoint,\n",
+        "    deployed_model_display_name=MODEL_DISPLAY_NAME,\n",
+        "    machine_type=\"n1-standard-8\",\n",
+        "    accelerator_type=\"NVIDIA_TESLA_P100\",\n",
+        "    accelerator_count=1,\n",
+        "    traffic_percentage=100,\n",
+        "    deploy_request_timeout=1200,\n",
+        "    sync=True,\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "c9fc560df0a9"
+      },
+      "source": [
+        "The previous cell, which deploys your model to the endpoint, can take a while to complete. If the previous cell times out before returning, your endpoint might still be successfully deployed to an endpoint. Check the [Cloud Console](https://console.cloud.google.com/vertex-ai/endpoints) to verify the results.\n",
+        "\n",
+        "You can also extend the time to wait for deployment by changing the `deploy_request_timeout` argument passed to `model.deploy()`."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "88a5304dfdc9"
+      },
+      "source": [
+        "## Get online predictions\n",
+        "\n",
+        "Finally, with your Stable Diffusion 1.5 model deployed to a Vertex AI endpoint, you can now get online predictions from it. Using the Vertex AI SDK, you only need a few lines of code to get an inference."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "0d6bc4aa34d6"
+      },
+      "outputs": [],
+      "source": [
+        "instances = [{\"prompt\": \"An examplePup dog with a baseball jersey.\"}]\n",
+        "response = endpoint.predict(instances=instances)\n",
+        "\n",
+        "with open(\"img5.jpg\", \"wb\") as g:\n",
+        "    g.write(base64.b64decode(response.predictions[0]))"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "65bafefda60c"
+      },
+      "outputs": [],
+      "source": [
+        "display.Image(\"img5.jpg\")"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "TpV-iwP9qw9c"
+      },
+      "source": [
+        "## Cleaning up\n",
+        "\n",
+        "To clean up all Google Cloud resources used in this project, you can [delete the Google Cloud\n",
+        "project](https://cloud.google.com/resource-manager/docs/creating-managing-projects#shutting_down_projects) you used for the tutorial.\n",
+        "\n",
+        "Otherwise, you can delete the individual resources you created in this tutorial:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "sx_vKniMq9ZX"
+      },
+      "outputs": [],
+      "source": [
+        "import os\n",
+        "\n",
+        "# Delete endpoint resource\n",
+        "endpoint.undeploy_all()\n",
+        "endpoint.delete()\n",
+        "\n",
+        "# Delete model resource\n",
+        "model.delete()\n",
+        "\n",
+        "# Delete Cloud Storage objects that were created\n",
+        "delete_bucket = False\n",
+        "if delete_bucket or os.getenv(\"IS_TESTING\"):\n",
+        "    ! gsutil -m rm -r $BUCKET_URI"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "name": "dreambooth_stablediffusion.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/notebook_template.ipynb

@@ -74,7 +74,9 @@
       "source": [
         "## Overview\n",
         "\n",
-        "{TODO: Include a paragraph or two explaining what this example demonstrates, who should be interested in it, and what you need to know before you get started.}"
+        "{TODO: Include a paragraph or two explaining what this example demonstrates, who should be interested in it, and what you need to know before you get started.}\n",
+        "\n",
+        "Learn more about [web-doc-title](linkback-to-webdoc-page). {TODO: if more than one primary feature, add tag/linkback for each one}"
       ]
     },
     {
@@ -160,7 +162,13 @@
       "outputs": [],
       "source": [
         "# Install the packages\n",
-        "! pip3 install --user --upgrade google-cloud-aiplatform"
+        "import os\n",
+        "\n",
+        "if not os.getenv(\"IS_TESTING\"):\n",
+        "    USER = \"--user\"\n",
+        "else:\n",
+        "    USER = \"\"\n",
+        "! pip3 install {USER} --upgrade google-cloud-aiplatform"
       ]
     },
     {

notebooks/notebook_template_review.py

@@ -12,6 +12,10 @@
         --errors-codes: A list of error codes to report errors. Otherwise, all errors are reported.
         --errors-csv: Report errors in CSV format
         
+        # options for automatic fixing
+        --fix: Automatic fix
+        --fix-codes: A list of fix codes to fix. Otherwise, all fix codes are enabled.
+        
         # index generatation
         --repo: Generate index in markdown format
         --web: Generate index in HTML format
@@ -19,6 +23,7 @@
         --desc: Add description to index
         --steps: Add steps to index
         --uses: Add "resources" used to index
+        --linkback: Add linkback to index
         
     Format of CSV file for notebooks to review:
     
@@ -61,6 +66,8 @@ parser.add_argument('--uses', dest='uses', action='store_true',
                     default=False, help='Output uses (resources)')
 parser.add_argument('--steps', dest='steps', action='store_true', 
                     default=False, help='Ouput steps')
+parser.add_argument('--linkback', dest='linkback', action='store_true', 
+                    default=False, help='Ouput linkback')
 parser.add_argument('--web', dest='web', action='store_true', 
                     default=False, help='Output format in HTML')
 parser.add_argument('--repo', dest='repo', action='store_true', 
@@ -109,13 +116,14 @@ class ErrorCode(Enum):
     #   Costs cell required
     #     Check for required Vertex and optional BQ and Dataflow
     ERROR_OVERVIEW_NOTFOUND = 10,
-    ERROR_OBJECTIVE_NOTFOUND = 11,
-    ERROR_OBJECTIVE_MISSING_DESC = 12,
-    ERROR_OBJECTIVE_MISSING_USES = 13,
-    ERROR_OBJECTIVE_MISSING_STEPS = 14,
-    ERROR_DATASET_NOTFOUND = 15,
-    ERROR_COSTS_NOTFOUND = 16,
-    ERROR_COSTS_MISSING = 17,
+    ERROR_LINKBACK_NOTFOUND = 11,
+    ERROR_OBJECTIVE_NOTFOUND = 12,
+    ERROR_OBJECTIVE_MISSING_DESC = 13,
+    ERROR_OBJECTIVE_MISSING_USES = 14,
+    ERROR_OBJECTIVE_MISSING_STEPS = 15,
+    ERROR_DATASET_NOTFOUND = 16,
+    ERROR_COSTS_NOTFOUND = 17,
+    ERROR_COSTS_MISSING = 18,
 
     # Installation cell
     #   Installation cell required
@@ -126,34 +134,34 @@ class ErrorCode(Enum):
     #   option {USER_FLAG} required
     #   installation code cell not match template
     #   all packages must be installed as a single pip3
-    ERROR_INSTALLATION_NOTFOUND = 18,
-    ERROR_INSTALLATION_HEADING = 19,
-    ERROR_INSTALLATION_CODE_NOTFOUND = 20,
-    ERROR_INSTALLATION_PIP3 = 21,
-    ERROR_INSTALLATION_QUIET = 22,
-    ERROR_INSTALLATION_USER_FLAG = 23,
-    ERROR_INSTALLATION_CODE_TEMPLATE = 24,
-    ERROR_INSTALLATION_SINGLE_PIP3 = 25,
+    ERROR_INSTALLATION_NOTFOUND = 19,
+    ERROR_INSTALLATION_HEADING = 20,
+    ERROR_INSTALLATION_CODE_NOTFOUND = 21,
+    ERROR_INSTALLATION_PIP3 = 22,
+    ERROR_INSTALLATION_QUIET = 23,
+    ERROR_INSTALLATION_USER_FLAG = 24,
+    ERROR_INSTALLATION_CODE_TEMPLATE = 25,
+    ERROR_INSTALLATION_SINGLE_PIP3 = 26,
 
     # Restart kernel cell
     #    Restart code cell required
     #    Restart code cell not found
-    ERROR_RESTART_NOTFOUND = 23,
-    ERROR_RESTART_CODE_NOTFOUND = 24,
+    ERROR_RESTART_NOTFOUND = 27,
+    ERROR_RESTART_CODE_NOTFOUND = 28,
 
     # Before you begin cell
     #    Before you begin cell required
     #    Before you begin cell incomplete
-    ERROR_BEFOREBEGIN_NOTFOUND = 25,
-    ERROR_BEFOREBEGIN_INCOMPLETE = 26,
+    ERROR_BEFOREBEGIN_NOTFOUND = 29,
+    ERROR_BEFOREBEGIN_INCOMPLETE = 30,
 
     # Set Project ID
     #    Set project ID cell required
     #    Set project ID code cell not found
     #    Set project ID not match template
-    ERROR_PROJECTID_NOTFOUND = 27,
-    ERROR_PROJECTID_CODE_NOTFOUND = 28,
-    ERROR_PROJECTID_TEMPLATE = 29,
+    ERROR_PROJECTID_NOTFOUND = 31,
+    ERROR_PROJECTID_CODE_NOTFOUND = 32,
+    ERROR_PROJECTID_TEMPLATE = 33,
 
     # Technical Writer Rules
     ERROR_TWRULE_TODO = 51,
@@ -182,7 +190,21 @@ def parse_dir(directory: str) -> int:
     """
     exit_code = 0
     
+    sorted_entries = []
     entries = os.scandir(directory)
+    for entry in entries:
+
+        inserted = False
+        for ix in range(len(sorted_entries)):
+            if entry.name < sorted_entries[ix].name:
+                sorted_entries.insert(ix, entry)
+                inserted = True
+                break
+        
+        if not inserted:
+            sorted_entries.append(entry)
+    
+    entries = sorted_entries
     for entry in entries:
         if entry.is_dir():
             if entry.name[0] == '.':
@@ -191,13 +213,65 @@ def parse_dir(directory: str) -> int:
                 continue
             exit_code += parse_dir(entry.path)
         elif entry.name.endswith('.ipynb'):
-            exit_code += parse_notebook(entry.path, tag=directory.split('/')[-1], linkback=None, rules=rules)
+            tag = directory.split('/')[-1]
+            if tag == 'automl':
+                tag = 'AutoML'
+            elif tag == 'bigquery_ml':
+                tag = 'BigQuery ML'
+            elif tag == 'custom':
+                tag = 'Vertex AI Training'
+            elif tag == 'experiments':
+                tag = 'Vertex AI Experiments'
+            elif tag == 'explainable_ai':
+                tag = 'Vertex Explainable AI'
+            elif tag == 'feature_store':
+                tag = 'Vertex AI Feature Store'
+            elif tag == 'matching_engine':
+                tag = 'Vertex AI Matching Engine'
+            elif tag == 'migration':
+                tag = 'CAIP to Vertex AI migration'
+            elif tag == 'ml_metadata':
+                tag = 'Vertex ML Metadata'
+            elif tag == 'model_evaluation':
+                tag = 'Vertex AI Model Evaluation'
+            elif tag == 'model_monitoring':
+                tag = 'Vertex AI Model Monitoring'
+            elif tag == 'model_registry':
+                tag = 'Vertex AI Model Registry'
+            elif tag == 'pipelines':
+                tag = 'Vertex AI Pipelines'
+            elif tag == 'prediction':
+                tag = 'Vertex AI Prediction'
+            elif tag == 'pytorch':
+                tag = 'Vertex AI Training'
+            elif tag == 'reduction_server':
+                tag = 'Vertex AI Reduction Server'
+            elif tag == 'sdk':
+                tag = 'Vertex AI SDK'
+            elif tag == 'structured_data':
+                tag = 'AutoML / BQML'
+            elif tag == 'tabnet':
+                tag = 'Vertex AI TabNet'
+            elif tag == 'tabular_workflows':
+                tag = 'AutoML Tabular Workflows'
+            elif tag == 'tensorboard':
+                tag = 'Vertex AI TensorBoard'
+            elif tag == 'training':
+                tag = 'Vertex AI Training'
+            elif tag == 'vizier':
+                tag = 'Vertex AI Vizier'
+                
+            # special case
+            if 'workbench' in directory:
+                tag = 'Vertex AI Workbench'
+                
+            exit_code += parse_notebook(entry.path, tags=[tag], linkback=None, rules=rules)
             
     return exit_code
 
 
 def parse_notebook(path: str,
-                   tag: str,
+                   tags: List,
                    linkback: str,
                    rules: List) -> int:
     """
@@ -205,8 +279,9 @@ def parse_notebook(path: str,
         and notebook authoring requirements.
         
             path: The path to the notebook.
-            tag: The associated tag
+            tags: The associated tags
             linkback: A link back to the web docs
+            rules: The cell rules to apply
             
         Returns the number of errors
     """
@@ -218,9 +293,20 @@ def parse_notebook(path: str,
     
     # Automatic Index Generation
     if objective.desc != '':
+        if overview.linkbacks:
+            linkbacks = overview.linkbacks
+        else:
+            if linkback:
+                linkbacks = [linkback]
+            else:
+                linkbacks = []
+
+        if overview.tags:
+            tags = overview.tags
+                
         add_index(path, 
-                  tag, 
-                  linkback,
+                  tags, 
+                  linkbacks,
                   title.title, 
                   objective.desc, 
                   objective.uses, 
@@ -512,9 +598,23 @@ class OverviewRule(NotebookRule):
         """
         Parse the overview cell
         """
+        self.linkbacks = []
+        self.tags = []
+        
         cell = notebook.get()
         if not cell['source'][0].startswith("## Overview"):
             return notebook.report_error(ErrorCode.ERROR_OVERVIEW_NOTFOUND, "Overview section not found")
+        
+        last_line = cell['source'][-1]
+        if last_line.startswith('Learn more about ['):
+            for more in last_line.split('[')[1:]:
+                tag = more.split(']')[0]
+                linkback = more.split('(')[1].split(')')[0]
+                self.tags.append(tag)
+                self.linkbacks.append(linkback)
+        else:
+            return notebook.report_error(ErrorCode.ERROR_LINKBACK_NOTFOUND, "Linkback missing in overview section")
+                
         return True
 
 
@@ -542,6 +642,10 @@ class ObjectiveRule(NotebookRule):
         in_steps = False
     
         for line in cell['source'][1:]:
+            # TOC anchor
+            if line.startswith('<a name='):
+                continue
+                
             if line.startswith('This tutorial uses'):
                 in_desc = False
                 in_steps = False
@@ -578,16 +682,39 @@ class ObjectiveRule(NotebookRule):
                         # check for italic font setting
                         if ch == '*' and sline[1] != ' ':
                             in_steps = False
+                        # special case
+                        elif sline.startswith('* Prediction Service'):
+                            in_steps = False
                         else:
                             self.steps += line
                     elif ch == '#':
                         in_steps = False
+
             
         if self.desc == '':
             ret = notebook.report_error(ErrorCode.ERROR_OBJECTIVE_MISSING_DESC, "Objective section missing desc")
         else:
             self.desc = self.desc.lstrip()
-            sentences = self.desc.split('.')
+            
+            bracket = False
+            paren = False
+            sentences = ""
+            for _ in range(len(self.desc)):
+                if self.desc[_] == '[':
+                    bracket = True
+                    continue
+                elif self.desc[_] == ']':
+                    bracket = False
+                    continue
+                elif self.desc[_] == '(':
+                    paren = True
+                elif self.desc[_] == ')':
+                    paren = False
+                    continue
+                    
+                if not paren:
+                    sentences += self.desc[_]
+            sentences = sentences.split('.')
             if len(sentences) > 1:
                 self.desc = sentences[0] + '.\n'
             if self.desc.startswith('In this tutorial, you learn') or self.desc.startswith('In this notebook, you learn'):
@@ -607,7 +734,7 @@ class ObjectiveRule(NotebookRule):
             ret = notebook.report_error(ErrorCode.ERROR_OBJECTIVE_MISSING_STEPS, "Objective section missing steps list")
             
         notebook.costs = self.costs
-        ret = True
+        return ret
 
 
 class RecommendationsRule(NotebookRule):
@@ -972,8 +1099,8 @@ class TextTWRule(TextRule):
 
 
 def add_index(path: str, 
-              tag: str, 
-              linkback: str,
+              tags: List, 
+              linkbacks: List,
               title : str, 
               desc: str, 
               uses: str, 
@@ -986,15 +1113,15 @@ def add_index(path: str,
     Add a discoverability index for this notebook
     
         path: The path to the notebook
-        tag: The tag (if any) for the notebook
+        tags: The tags (if any) for the notebook
         title: The H1 title for the notebook
-        desc:
-        uses:
-        steps:
-        git_link:
-        colab_link:
-        workbench_link:
-        linkback:
+        desc: The notebook description
+        uses: The resources/services used by the notebook
+        steps: The steps specified by the notebook
+        git_link: The link to the notebook in the git repo
+        colab_link: Link to launch notebook in Colab
+        workbench_link: Link to launch notebook in Workbench
+        linkbacks: The linkbacks per tag
     """
     global last_tag
     
@@ -1004,43 +1131,65 @@ def add_index(path: str,
     title = title.split(':')[-1].strip()
     title = title[0].upper() + title[1:]
     if args.web:
-        title = title.replace('`', '')
+        title = replace_cl(title.replace('`', ''))
         
         print('    <tr>')
         print('        <td>')
-        tags = tag.split(',')
         for tag in tags:
+            tag = replace_cl(tag)
             print(f'            {tag.strip()}<br/>\n')
         print('        </td>')
         print('        <td>')
-        print(f'            {title}<br/>\n')
+        print(f'            <b>{title}</b>. ')
         if args.desc:
-            desc = desc.replace('`', '')
-            print(f'            {desc}<br/>\n')
-        if linkback:
-            text = ''
-            for tag in tags:
-                text += tag.strip() + ' '
-                
-            print(f'            Learn more about <a src="https://cloud.google.com/{linkback}">{text}</a><br/>\n')
+            desc = replace_cl(desc.replace('`', ''))
+            print('<br/>')
+            print(f'            {desc}\n')
+            
+            
+        if args.linkback and linkbacks:
+            num = len(tags)
+            for _ in range(num):
+                if linkbacks[_].startswith("vertex-ai"):
+                    print(f' Learn more about <a href="https://cloud.google.com/{linkbacks[_]}" target="_blank">{replace_cl(tags[_])}</a>.\n')
+                else:
+                    print(f' Learn more about <a href="{linkbacks[_]}" target="_blank">{replace_cl(tags[_])}</a>.\n')
+                    
+        if args.steps:
+            print("<devsite-expandable>\n")
+            print('  <p class="showalways">Tutorial steps</p>\n')
+            print('  <ul>\n')
+            
+            if ":" in steps:
+                steps = steps.split(':')[1].replace('*', '').replace('-', '').strip().split('\n')
+            else:
+                steps = []
+              
+            for step in steps:
+                print(f'    <li>{replace_cl(step)}</li>\n')
+            print('  </ul>\n')
+            print("</devsite-expandable>\n")
+                    
         print('        </td>')
         print('        <td>')
         if colab_link:
-            print(f'            <a src="{colab_link}">Colab</a><br/>\n')
+            print(f'            <a href="{colab_link}" target="_blank" class="external" track-type="notebookTutorial" track-name="colabLink">Colab</a><br/>\n')
         if git_link:
-            print(f'            <a src="{git_link}">GitHub</a><br/>\n')
+            print(f'            <a href="{git_link}" target="_blank" class="external" track-type="notebookTutorial" track-name="gitHubLink">GitHub</a><br/>\n')
         if workbench_link:
-            print(f'            <a src="{workbench_link}">Vertex AI Workbench</a><br/>\n')
+            print(f'            <a href="{workbench_link}" target="_blank" class="external" track-type="notebookTutorial" track-name="workbenchLink">Vertex AI Workbench</a><br/>\n')
         print('        </td>')
         print('    </tr>\n')
     elif args.repo:
-        tags = tag.split(',')
-        if tags != last_tag and tag != '':
-            last_tag = tags
-            flat_list = ''
-            for item in tags:
-                flat_list += item.replace("'", '') + ' '
-            print(f"\n### {flat_list}\n")
+        try:
+            if tags != last_tag and tag != '':
+                last_tag = tags
+                flat_list = ''
+                for item in tags:
+                    flat_list += item.replace("'", '') + ' '
+                print(f"\n### {flat_list}\n")
+        except:
+            pass
         print(f"\n[{title}]({git_link})\n")
     
         print("```")
@@ -1052,7 +1201,84 @@ def add_index(path: str,
 
         if args.steps:
             print(steps.rstrip() + '\n')
+            
         print("```\n")
+            
+        if args.linkback and linkbacks:
+            num = len(tags)
+            for _ in range(num):
+                if linkbacks[_].startswith("vertex-ai"):
+                    print(f'&nbsp;&nbsp;&nbsp;Learn more about [{tags[_]}]({linkbacks[_]}).\n')
+                else:
+                    print(f'&nbsp;&nbsp;&nbsp;Learn more about [{tags[_]}]({linkbacks[_]}).\n')
+
+def replace_cl(text : str ) -> str:
+    '''
+    Replace product names with CL substitution variables
+    '''
+    substitutions = {
+        #'AutoML Tabular Workflow': '{{automl_name}} Tabular Workflow',
+        #'AutoML Tables': '{{automl_tables_name}}',
+        #'AutoML Tabular': '{{automl_tables_name}}',
+        #'AutoML Vision': '{automl_vision_name}}',
+        #'AutoML Image': '{automl_vision_name}}',
+        'AutoML': '{{automl_name}}',
+        
+        'BigQuery ML': '{{bigqueryml_name}}',
+        'BQML': '{{bigqueryml_name}}',
+        'BigQuery': '{{bigquery_name}}',
+        'BQ': '{{bigquery_name}}',
+        
+        'Vertex Dataset': '{{vertex_ai_name}} Dataset',
+        'Vertex Model': '{{vertex_ai_name}} Model',
+        'Vertex Endpoint': '{{vertex_ai_name}} Endpoint',
+        'Vertex Model Registry': '{{vertex_model_registry_name}}',
+        'Vertex AI Model Registry': '{{vertex_model_registry_name}}',
+        'Vertex Training': '{{vertex_training_name}}',
+        'Vertex AI Training': '{{vertex_training_name}}',
+        'Vertex Prediction': '{{vertex_prediction_name}}',
+        'Vertex AI Prediction': '{{vertex_prediction_name}}',
+        'Vertex TensorBoard': '{{vertex_tensorboard_name}}',
+        'Vertex AI TensorBoard': '{{vertex_tensorboard_name}}',
+        'TensorBoard': '{{vertex_tensorboard_name}}',
+        'Tensorboard': '{{vertex_tensorboard_name}}',
+        'Vertex ML Metadata': '{{vertex_metadata_name}}',
+        'Vertex Pipelines': '{{vertex_pipelines_name}}',
+        'Vertex AI Pipelines': '{{vertex_pipelines_name}}',
+        'Vertex AI Data Labeling': '{{vertex_data_labeling_name}}',
+        'Vertex AI Experiments': '{{vertex_experiments_name}}',
+        'Vertex Experiments': '{{vertex_experiments_name}}',
+        'Vertex AI Matching Engine': '{{vertex_matching_engine_name}}',
+        'Vertex Matching Engine': '{{vertex_matching_engine_name}}',
+        'Vertex Model Monitoring': '{{vertex_model_monitoring_name}}',
+        'Vertex AI Model Monitoring': '{{vertex_model_monitoring_name}}',
+        'Vertex Feature Store': '{{vertex_featurestore_name}}',
+        'Vertex AI Feature Store': '{{vertex_featurestore_name}}',
+        'Vertex Vizier': '{{vertex_vizier_name}}',
+        'Vertex AI Vizier': '{{vertex_vizier_name}}',
+        'Vertex Explainable AI': '{{vertex_xai_name}}',
+        'NAS': '{{vertex_nas_name}',
+        'Vertex AI Neural Architectural Search': '{{vertex_nas_name}}',
+        'Vertex Workbench': '{{vertex_workbench_name}}',
+        'Vertex AI Workbench': '{{vertex_workbench_name}}',
+        'Vertex AI Edge Manager': '{{vertex_edge_manager_name}}',
+        'Vertex SDK': '{{vertex_sdk_name}}',
+        'Vertex AI SDK': '{{vertex_sdk_name}}',
+        'Vertex AI': '{{vertex_ai_name}}',
+        
+        'Cloud Storage': '{{storage_name}}',
+        'GCS': '{{storage_name}}',
+        'GCP': '{{gcp_name}}',
+        'TensorFlow Enterprise': '{{tf4gcp_name}}',
+        'TensorFlow': '{{tensorflow_name}}',
+    }
+    
+    for key, value in substitutions.items():
+        if key in text:
+            text = text.replace(key, value)
+            
+    return text
+
 
 
 # Instantiate the rules
@@ -1084,21 +1310,32 @@ rules = [ copyright, notices, title, links, testenv, table, overview, objective,
 ]
 
 if args.web:
+    print('<style>')
+    print('table, th, td {')
+    print('  border: 1px solid black;')
+    print('  padding-left:10px')
+    print('}')
+    print('</style>')
     print('<table>')
-    print('    <th>Vertex AI Feature</th>')
-    print('    <th>Description</th>')
-    print('    <th>Open in</th>')
+    print('    <thead>')
+    print('        <tr>')
+    print('            <th width="180px">Services</th>')
+    print('            <th>Description</th>')
+    print('            <th width="80px">Open in</th>')
+    print('        </tr>')
+    print('    </thead>')
+    print('    <tbody class="list">')
 
 if args.notebook_dir:
     if not os.path.isdir(args.notebook_dir):
-        print("Error: not a directory:", args.notebook_dir)
+        print(f"Error: not a directory: {args.notebook_dir}", file=sys.stderr)
         exit(1)
     exit_code = parse_dir(args.notebook_dir)
 elif args.notebook:
     if not os.path.isfile(args.notebook):
-        print("Error: not a notebook:", args.notebook)
+        print(f"Error: not a notebook: {args.notebook}", file=sys.stderr)
         exit(1)
-    exit_code = parse_notebook(args.notebook, tag='', linkback=None, rules=rules)
+    exit_code = parse_notebook(args.notebook, tags=[], linkback=None, rules=rules)
 elif args.notebook_file:
     if not os.path.isfile(args.notebook_file):
         print("Error: file does not exist", args.notebook_file)
@@ -1111,18 +1348,19 @@ elif args.notebook_file:
                 if heading:
                     heading = False
                 else:
-                    tag = row[0]
+                    tags = row[0].split(',')
                     notebook = row[1]
                     try:
                         linkback = row[2]
                     except:
                         linkback = None
-                    exit_code += parse_notebook(notebook, tag=tag, linkback=linkback, rules=rules)
+                    exit_code += parse_notebook(notebook, tags=tags, linkback=linkback, rules=rules)
 else:
-    print("Error: must specify a directory or notebook")
+    print("Error: must specify a directory or notebook", file=sys.stderr)
     exit(1)
 
 if args.web:
+    print('    </tbody>\n')
     print('</table>\n')
 
 exit(exit_code)

notebooks/notebook_template_writing_guidelines.md

@@ -23,7 +23,62 @@
     - Incorrect examples: "Let's update the field", "We'll update the field", "The user should update the field"
 - **Googlers**: Please follow our [branding guidelines](http://goto/cloud-branding).
 
-### Code
+
+## Authoring guidelines
+
+### Focus
+
+Notebooks for official are expected to be narrow focused, which highlight a subset of features of a Vertex AI product/service.
+The product/feature is to be highlighted in the Overview section. For example:
+
+```
+This tutorial demonstrates using Vertex AI Training to train an XGBoost model using a XGBoost pre-built training container.
+```
+
+In the above example, the Vertex AI product/service is `Vertex AI Training` and the feature is `XGBoost pre-built training container`.
+
+### Scope
+
+Notebooks for official are expected to be narrow in scope, without extra extraneous steps. For example, if the notebook is about training, we discourage ending the notebook with deploying the model and doing an online/batch prediction. On the later, we recommend a separate notebook about prediction that uses a pretrained model.
+
+#### Training
+
+Notebooks for training should be constructed as follows:
+
+1. If the training script(s) are small, embed them in the notebook and use %writefile to store them locally.
+2. If the training script(s) are large, store them in our public bucket: gs://cloud-samples-data/vertex-ai/dataset-management/script, and use !wgets to retrieve and store the script locally.
+3. Train the model using the Vertex AI SDK methods for custom training.
+4. Preferrable have the service upload the trained model to the Vertex AI Model Registry.
+4. Have the script do an evaluation.
+5. Retrieve the evaluation metrics and attach them as an artifact to the corresponding entry in the Model Registry.
+6. Optionally, download the model artifacts and test locally -- i.e., make a local prediction request.
+
+#### Evaluation
+
+Notebooks for evaluation should be constructed as follows:
+
+1. Use a pretrained model from a public repository.
+2. Upload the pretrained model to the Vertex AI Model Registry.
+3. Perform a model evaluation.
+4. Review the model evaluation.
+4. Attach the model evaluation to the corresponding entry in the Model Registry.
+
+#### Prediction
+
+Notebooks for prediction should be constructed as follows:
+
+1. Use a pretrained model from a public repository.
+2. If relevant, attach a serving function to the model artifacts.
+3. Upload the pretrained model to the Vertex AI Model Registry.
+4. For online:<br/>
+    A. Deploy the model.<br/>
+    B. Perform an online prediction.</br>
+    C. Review the result.
+5. For batch:<br/>
+    A. Perform a batch prediction.</br/>
+    B. Review the result.
+    
+## Code
 
 - Put all your installs and imports in a setup section.
 - Save the notebook with the Table of Contents open.
@@ -31,7 +86,7 @@
 - Follow the [Google Python Style guide](https://github.com/google/styleguide/blob/gh-pages/pyguide.md) and write readable code.
 - Keep cells small (max ~20 lines).
 
-## TensorFlow code style
+### TensorFlow code style
 
 Use the highest level API that gets the job done (unless the goal is to demonstrate the low level API). For example, when using Tensorflow:
 

notebooks/official.csv

@@ -1,70 +0,0 @@
-tag,notebook,doc
-"AutoML, Text data",official/automl/automl-text-classification.ipynb,vertex-ai/docs/text-data/classification/train-model
-"AutoML, Text data",official/automl/sdk_automl_text_entity_extraction_online.ipynb,
-"AutoML, Text data",official/automl/sdk_automl_text_sentiment_analysis_online.ipynb,
-"AutoML, Tabular data",official/automl/sdk_automl_tabular_forecasting_batch.ipynb,vertex-ai/docs/tabular-data/forecasting/tutorials-samples
-"AutoML, Tabular Data",official/automl/automl_tabular_on_vertex_pipelines.ipynb,vertex-ai/docs/tabular-data/tabular-workflows/e2e-automl
-"AutoML, Tabular Data",official/automl/sdk_automl_tabular_regression_batch_bq.ipynb,
-"AutoML, Tabular Data",official/automl/sdk_automl_tabular_regression_batch_bq.ipynb,
-"AutoML, Forecasting",official/automl/automl_forecasting_bqml_arima_plus_comparison.ipynb,vertex-ai/docs/tabular-data/forecasting-arima/overview
-"AutoML, Forecasting",official/automl/sdk_automl_tabular_forecasting_batch.ipynb,
-"AutoML, Image data",official/automl/sdk_automl_text_sentiment_analysis_online.ipynb,
-"AutoML, Video data",official/automl/sdk_automl_text_sentiment_analysis_online.ipynb,
-"AutoML, Video data",official/automl/sdk_automl_video_classification_batch.ipynb,
-"AutoML, Video data",official/automl/sdk_automl_video_object_tracking_batch.ipynb,
-"AutoML, Video data",official/sdk/SDK_AutoML_Video_Classification.ipynb,
-"BigQuery, Vertex AI Workbench",official/workbench/exploratory_data_analysis/explore_data_in_bigquery_with_workbench.ipynb,
-"BigQuery ML, Vertex AI Model Registry, Batch prediction",official/model_registry/bqml_vertexai_model_registry.ipynb,
-"BigQuery ML, Vertex AI Model Registry, Online prediction",official/bigquery_ml/bqml-online-prediction.ipynb,
-"BigQuery ML",official/structured_data/rapid_prototyping_bqml_automl.ipynb,
-Custom Training,official/custom/sdk-custom-image-classification-batch.ipynb,
-Custom Training,official/custom/sdk-custom-image-classification-online.ipynb,
-Custom Training,official/custom/SDK_Custom_Container_Prediction.ipynb,
-"Custom Training, BiqQuery dataset",official/custom/custom-tabular-bq-managed-dataset.ipynb,
-"Custom Training, TensorBoard",official/custom/custom-tabular-bq-managed-dataset.ipynb,
-"Custom Training, TensorBoard",official/tensorboard/tensorboard_custom_training_with_custom_container.ipynb,
-"Custom Training, TensorBoard",official/tensorboard/tensorboard_custom_training_with_prebuilt_container.ipynb
-"Custom Training, Managed dataset",official/sdk/SDK_Custom_Training_Python_Package_Managed_Text_Dataset_Tensorflow_Serving_Container.ipynb,
-"Custom Training, Distributed",official/training/multi_node_ddp_gloo_vertex_training_with_custom_container.ipynb,
-"Custom Training, Distributed",official/training/multi_node_ddp_nccl_vertex_training_with_custom_container.ipynb
-Vertex AI Experiments,official/experiments/comparing_pipeline_runs.ipynb,
-Vertex AI Experiments,official/experiments/build_model_experimentation_lineage_with_prebuild_code.ipynb,
-Vertex AI Experiments,official/experiments/comparing_local_trained_models.ipynb,
-"Vertex Explainable AI, Tabular data",official/explainable_ai/sdk_automl_tabular_binary_classification_batch_explain.ipynb,vertex-ai/docs/explainable-ai/overview
-"Vertex Explainable AI, Tabular data",official/explainable_ai/sdk_automl_tabular_classification_online_explain.ipynb,vertex-ai/docs/explainable-ai/overview
-"Vertex Explainable AI, Image data",official/explainable_ai/sdk_custom_image_classification_batch_explain.ipynb,vertex-ai/docs/explainable-ai/overview
-"Vertex Explainable AI, Tabular data",official/explainable_ai/sdk_custom_tabular_regression_batch_explain.ipynb,vertex-ai/docs/explainable-ai/overview
-"Vertex Explainable AI, Tabular data",official/explainable_ai/sdk_custom_tabular_regression_online_explain.ipynb,vertex-ai/docs/explainable-ai/overview
-Vertex ML Metadata,official/ml_metadata/sdk-metric-parameter-tracking-for-custom-jobs.ipynb,
-"Vertex Explainable AI, Image data",official/explainable_ai/sdk_custom_image_classification_online_explain.ipynb,vertex-ai/docs/explainable-ai/overview
-Vertex AI Feature Store,official/feature_store/sdk-feature-store.ipynb,
-Vertex AI Feature Store,official/feature_store/sdk-feature-store-pandas.ipynb,
-Vertex AI Matching Engine,official/matching_engine/sdk_matching_engine_for_indexing.ipynb,
-Vertex ML Metadata,official/ml_metadata/sdk-metric-parameter-tracking-for-custom-jobs.ipynb,
-Vertex ML Metadata,official/ml_metadata/sdk-metric-parameter-tracking-for-locally-trained-models.ipynb,
-"Vertex ML Metadata, Vertex AI Pipelines",official/ml_metadata/sdk-metric-parameter-tracking-for-locally-trained-models.ipynb,
-"Vertex AI Model Evaluation, AutoML",official/model_evaluation/automl_tabular_classification_model_evaluation.ipynb,
-"Vertex AI Model Evaluation, AutoML",official/model_evaluation/automl_tabular_regression_model_evaluation.ipynb,
-"Vertex AI Model Evaluation, AutoML",official/model_evaluation/automl_text_classification_model_evaluation.ipynb,
-"Vertex AI Model Evaluation, AutoML",official/model_evaluation/automl_video_classification_model_evaluation.ipynb,
-"Vertex AI Model Evaluation, Custom Training",official/model_evaluation/custom_tabular_regression_model_evaluation.ipynb,
-Model Monitoring,official/model_monitoring/model_monitoring.ipynb,
-Vertex AI Pipelines,official/pipelines/pipelines_intro_kfp.ipynb,
-Vertex AI Pipelines,official/pipelines/control_flow_kfp.ipynb,
-Vertex AI Pipelines,official/pipelines/metrics_viz_run_compare_kfp.ipynb,
-Vertex AI Pipelines,official/pipelines/lightweight_functions_component_io_kfp.ipynb,
-"Vertex AI Pipelines Image data",official/pipelines/google_cloud_pipeline_components_automl_images.ipynb,
-"Vertex AI Pipelines, Tabular data",official/pipelines/automl_tabular_classification_beans.ipynb,
-"Vertex AI Pipelines, Tabular data",official/pipelines/google_cloud_pipeline_components_automl_tabular.ipynb,
-"Vertex AI Pipelines, Tabular data",official/pipelines/google_cloud_pipeline_components_dataproc_tabular.ipynb,
-"Vertex AI Pipelines, Text data",official/pipelines/google_cloud_pipeline_components_automl_text.ipynb,
-"Vertex AI Pipelines, Text data",official/pipelines/google_cloud_pipeline_components_bqml_text.ipynb,
-Vertex AI Pipelines,official/pipelines/custom_model_training_and_batch_prediction.ipynb,
-Vertex AI Pipelines,official/pipelines/google_cloud_pipeline_components_model_train_upload_deploy.ipynb,
-Vertex AI Pipelines,official/pipelines/google_cloud_pipeline_components_model_upload_predict_evaluate.ipynb,
-"Vertex AI Training, Reduction Server, PyTorch",official/reduction_server/pytorch_distributed_training_reduction_server.ipynb,
-"Tabular Workflows, Vertex AI TabNet",official/tabnet/tabnet_vertex_tutorial.ipynb,
-"Tabular Workflows, Vertex AI TabNet, Vertex Explainablee AI",official/tabnet/ai-explanations-tabnet-algorithm.ipynb,
-"Tabular Workflows, Vertex AI TabNet, Vertex AI Pipelines",official/tabular_workflows/tabnet_on_vertex_pipelines.ipynb,
-"Tabular Workflows, Vertex AI Wide and Deep",official/tabular_workflows/wide_and_deep_on_vertex_pipelines.ipynb,
-Vertex AI Vizier,official/vizier/gapic-vizier-multi-objective-optimization.ipynb,vertex-ai/docs/vizier/using-vizier

notebooks/official/CODEOWNERS

@@ -29,17 +29,15 @@
 /pipelines/google_cloud_pipelines_dataproc_tabular @inardini
 /automl/automl_forecasting_bqml_arima_plus_comparison.ipynb @TheMichaelHu
 /automl/automl_tabular_on_vertex_pipelines.ipynb @helinwang
-/custom/custom_training_tensorboard_profiler.ipynb @itseric
+/custom/custom_training_tensorboard_profiler.ipynb @gericdong
+/custom/get_started_with_vertex_endpoint_and_shared_vm.ipynb @andrewferlitsch
 /workbench/spark/spark_sample_notebook.ipynb @bradmiro
 /workbench/spark/spark_ml.ipynb @bradmiro
 /model_registry/bqml_vertexai_model_registry.ipynb @soheilazangeneh
-/workbench/exploratory_data_analysis/explore_data_in_bigquery_with_workbench.ipynb @alokpattani
 /model_evaluation/automl_tabular_classification_model_evaluation.ipynb @soheilazangeneh
 /model_evaluation/automl_tabular_regression_model_evaluation.ipynb @soheilazangeneh
-/tabular_workflows/tabnet_on_vertex_pipelines.ipynb @sakagarwal
-/tabular_workflows/wide_and_deep_on_vertex_pipelines.ipynb @sakagarwal
+/tabular_workflows/prophet_on_vertex_pipelines.ipynb @TheMichaelHu
 /model_evaluation/custom_tabular_classification_model_evaluation.ipynb @soheilazangeneh
 /sdk/SDK_FBProphet_Forecasting_Online.ipynb @brianchunkang
-/automl/sdk_automl_forecasting_hierarchical_batch.ipynb @ivanmkc
-/prediction/custom_batch_prediction_feature_filter.ipynb @soheilazangeneh
-/feature_store/feature_store_streaming_ingestion_sdk.ipynb @soheilazangeneh
+/pipelines/Train_tabular_models_with_many_frameworks_and_import_to_Vertex_AI_using_Pipelines @Ark-kun
+/experiments/get_started_with_vertex_experiments_autologging.ipynb @inardini

notebooks/official/README.md

@@ -4,441 +4,3 @@ The official notebooks are a collection of curated and non-curated notebooks aut
 
 The official notebooks are organized by Google Cloud Vertex AI services.
 
-## Manifest of Curated Notebooks
-
-### AutoML  Text data 
-
-
-[Create, train, and deploy an AutoML text classification model](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl-text-classification.ipynb)
-
-Learn how to use `AutoML` to train a text classification model.
-
-The steps performed include:
-
-* Create a `Vertex AI Dataset`.
-* Train an `AutoML` text classification `Model` resource.
-* Obtain the evaluation metrics for the `Model` resource.
-* Create an `Endpoint` resource.
-* Deploy the `Model` resource to the `Endpoint` resource.
-* Make an online prediction
-* Make a batch prediction
-
-### AutoML  Tabular data 
-
-
-[AutoML tabular forecasting model for batch prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/sdk_automl_tabular_forecasting_batch.ipynb)
-
-Learn how to create an `AutoML` tabular forecasting model from a Python script, and then do a batch prediction using the Vertex AI SDK.
-
-The steps performed include:
-
-- Create a `Vertex AI Dataset` resource.
-- Train an `AutoML` tabular forecasting `Model` resource.
-- Obtain the evaluation metrics for the `Model` resource.
-- Make a batch prediction.
-
-### BigQuery ML  Vertex AI Model Registry  Batch prediction 
-
-
-[Deploy BiqQuery ML Model on Vertex AI Model Registry and make predictions](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/model-registry/bqml-vertexai-model-registry.ipynb)
-
-Learn how to use `Vertex AI Model Registry` with `BigQuery ML` and make batch predictions:
-
-The steps performed include:
-
-- Train a model with `BigQuery ML`
-- Upload the model to `Vertex AI Model Registry` 
-- Create a `Vertex AI Endpoint` resource
-- Deploy the `Model` resource to the `Endpoint` resource
-- Make `prediction` requests to the model endpoint
-- Run `batch prediction` job on the `Model` resource 
-
-
-### BigQuery ML  Vertex AI Model Registry  Online prediction 
-
-
-[Online prediction with BigQuery ML](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/bigquery_ml/bqml-online-prediction.ipynb)
-
-Learn how to train and deploy a churn prediction model for real-time inference, with the data in BigQuery and model trained using BigQuery ML, registered to Vertex AI Model Registry, and deployed to an endpoint on Vertex AI for online predictions.
-
-The steps performed include:
-
-- Using Python & SQL to query the public data in BigQuery
-- Preparing the data for modeling
-- Training a classification model using BigQuery ML and registering it to Vertex AI Model Registry
-- Inspecting the model on Vertex AI Model Registry
-- Deploying the model to an endpoint on Vertex AI
-- Making sample online predictions to the model endpoint
-
-
-### Custom Training 
-
-
-[Custom training and batch prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/custom/sdk-custom-image-classification-batch.ipynb)
-
-Learn to use `Vertex AI Training` to create a custom trained model and use `Vertex AI Batch Prediction` to do a batch prediction on the trained model.
-
-The steps performed include:
-
-- Create a `Vertex AI` custom job for training a TensorFlow model.
-- Upload the trained model artifacts as a `Model` resource.
-- Make a batch prediction.
-
-[Custom training and online prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/custom/sdk-custom-image-classification-online.ipynb)
-
-Learn to use `Vertex AI Training` to create a custom-trained model from a Python script in a Docker container, and learn to use `Vertex AI Prediction` to do a prediction on the deployed model by sending data.
-
-The steps performed include:
-
-- Create a `Vertex AI` custom job for training a TensorFlow model.
-- Upload the trained model artifacts to a `Model` resource.
-- Create a serving `Endpoint` resource.
-- Deploy the `Model` resource to a serving `Endpoint` resource.
-- Make a prediction.
-- Undeploy the `Model` resource.
-
-### Tabular Data 
-
-
-[Compare Vertex AI Forecasting and BigQuery ML ARIMA_PLUS](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_forecasting_bqml_arima_plus_comparison.ipynb)
-
-Learn how to create an BQML ARIMA_PLUS model using a training [Vertex AI Pipeline](https://cloud.
-
-The steps performed are:
-
-- Train the BQML ARIMA_PLUS model.
-- View BQML model evaluation.
-- Make a batch prediction with the BQML model.
-- Create a Vertex AI `Dataset` resource.
-- Train the Vertex AI Forecasting model.
-- View the Model evaluation.
-- Make a batch prediction with the Model.
-
-
-### AutoML  Tabular Data 
-
-
-[AutoML Tabular Pipeline](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_tabular_on_vertex_pipelines.ipynb)
-
-Learn how to create two regression models using [Vertex Pipelines](https://cloud.
-
-The steps performed are:
-
-- Create a training pipeline that reduces the search space from the default to save time.
-- Create a training pipeline that reuses the architecture search results from the previous pipeline to save time.
-
-### Vertex AI Experiments 
-
-
-[Compare pipeline runs with Vertex AI Experiments](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/experiments/comparing_pipeline_runs.ipynb)
-
-Learn how to use `Vertex AI Experiments` to log a pipeline job and compare different pipeline jobs.
-
-
-
-[Build Vertex AI Experiment lineage for custom training](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/experiments/build_model_experimentation_lineage_with_prebuild_code.ipynb)
-
-Learn how to integrate preprocessing code in a Vertex AI experiments.
-
-
-
-[Track parameters and metrics for locally trained models](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/experiments/comparing_local_trained_models.ipynb)
-
-Learn how to use Vertex AI Experiments to compare and evaluate model experiments.
-
-The steps performed include:
-
-- log the model parameters
-- log the loss and metrics on every epoch to TensorBoard
-- log the evaluation metrics
-
-
-### Vertex AI Feature Store 
-
-
-[Online and Batch predictions using Vertex AI Feature Store](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/feature_store/sdk-feature-store.ipynb)
-
-Learn how to use `Vertex AI Feature Store` to import feature data, and to access the feature data for both online serving and offline tasks, such as training.
-
-The steps performed include:
-
-- Create featurestore, entity type, and feature resources.
-- Import feature data into `Vertex AI Feature Store` resource.
-- Serve online prediction requests using the imported features.
-- Access imported features in offline jobs, such as training jobs.
-
-### Matching Engine 
-
-
-[Create Vertex AI Matching Engine index](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/matching_engine/sdk_matching_engine_for_indexing.ipynb)
-
-Learn how to create Approximate Nearest Neighbor (ANN) Index, query against indexes, and validate the performance of the index.
-
-The steps performed include:
-
-* Create ANN Index and Brute Force Index
-* Create an IndexEndpoint with VPC Network
-* Deploy ANN Index and Brute Force Index
-* Perform online query
-* Compute recall
-
-
-### Model Monitoring 
-
-
-[Vertex AI Model Monitoring with Explainable AI Feature Attributions](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/model_monitoring/model_monitoring.ipynb)
-
-Learn to use the `Vertex AI Model Monitoring` service to detect drift and anomalies in prediction requests from a deployed `Vertex AI Model` resource.
-
-The steps performed include:
-
-- Upload a pre-trained model as a `Vertex AI Model` resource.
-- Create an `Vertex AI Endpoint` resource.
-- Deploy the `Model` resource to the `Endpoint` resource.
-- Configure the `Endpoint` resource for model monitoring.
-- Initialize the baseline distribution for model monitoring.
-- Generate synthetic prediction requests.
-- Understand how to interpret the statistics, visualizations, other data reported by the model monitoring feature.
-
-### Vertex AI Pipelines 
-
-
-[Lightweight Python function-based components, and component I/O](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/pipelines/lightweight_functions_component_io_kfp.ipynb)
-
-Learn to use the KFP SDK to build lightweight Python function-based components, and then you learn to use `Vertex AI Pipelines` to execute the pipeline.
-
-The steps performed include:
-
-- Build Python function-based KFP components.
-- Construct a KFP pipeline.
-- Pass *Artifacts* and *parameters* between components, both by path reference and by value.
-- Use the `kfp.dsl.importer` method.
-- Compile the KFP pipeline.
-- Execute the KFP pipeline using `Vertex AI Pipelines`
-
-### Vertex AI Pipelines Image data 
-
-
-[AutoML image classification pipelines using google-cloud-pipeline-components](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/pipelines/google_cloud_pipeline_components_automl_images.ipynb)
-
-Learn how to use `Vertex AI Pipelines` and `Google Cloud Pipeline Components` to build an `AutoML` image classification model.
-
-The steps performed include:
-
-- Create a KFP pipeline:
-    - Create a `Dataset` resource.
-    - Train an AutoML image classification `Model` resource.
-    - Create an `Endpoint` resource.
-    - Deploys the `Model` resource to the `Endpoint` resource.
-- Compile the KFP pipeline.
-- Execute the KFP pipeline using `Vertex AI Pipelines`
-
-
-
-### Vertex AI Pipelines  Tabular data 
-
-
-[AutoML Tabular pipelines using google-cloud-pipeline-components](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/pipelines/automl_tabular_classification_beans.ipynb)
-
-Learn to use `Vertex AI Pipelines` and `Google Cloud Pipeline Components` to build an `AutoML` tabular classification model.
-
-The steps performed include:
-
-- Create a KFP pipeline:
-    - Create a `Dataset` resource.
-    - Train an AutoML tabular classification `Model` resource.
-    - Create an `Endpoint` resource.
-    - Deploys the `Model` resource to the `Endpoint` resource.
-- Compile the KFP pipeline.
-- Execute the KFP pipeline using `Vertex AI Pipelines`
-
-
-
-[AutoML tabular regression pipelines using google-cloud-pipeline-components](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/pipelines/google_cloud_pipeline_components_automl_tabular.ipynb)
-
-Learn to use `Vertex AI Pipelines` and `Google Cloud Pipeline Components` to build an `AutoML` tabular regression model.
-
-The steps performed include:
-
-- Create a KFP pipeline:
-    - Create a `Dataset` resource.
-    - Train an AutoML tabular regression `Model` resource.
-    - Create an `Endpoint` resource.
-    - Deploys the `Model` resource to the `Endpoint` resource.
-- Compile the KFP pipeline.
-- Execute the KFP pipeline using `Vertex AI Pipelines`
-
-
-
-### Vertex AI Pipelines  Text data 
-
-
-[AutoML text classification pipelines using google-cloud-pipeline-components](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/pipelines/google_cloud_pipeline_components_automl_text.ipynb)
-
-Learn to use `Vertex AI Pipelines` and `Google Cloud Pipeline Components` to build an `AutoML` text classification model.
-
-The steps performed include:
-
-- Create a KFP pipeline:
-    - Create a `Dataset` resource.
-    - Train an AutoML text classification `Model` resource.
-    - Create an `Endpoint` resource.
-    - Deploys the `Model` resource to the `Endpoint` resource.
-- Compile the KFP pipeline.
-- Execute the KFP pipeline using `Vertex AI Pipelines`
-
-
-
-### Vertex AI Pipelines 
-
-
-[Custom training with pre-built Google Cloud Pipeline Components](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/pipelines/custom_model_training_and_batch_prediction.ipynb)
-
-Learn to use `Vertex AI Pipelines` and `Google Cloud Pipeline Components` to build a custom model.
-
-The steps performed include:
-
-- Create a KFP pipeline:
-    - Train a custom model.
-    - Upload the trained model as a `Model` resource.
-    - Create an `Endpoint` resource.
-    - Deploy the `Model` resource to the `Endpoint` resource.
-    - Make a batch prediction request.
-
-
-
-[Pipeline control structures using the KFP SDK](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/pipelines/control_flow_kfp.ipynb)
-
-Learn how to use the KFP SDK to build pipelines that use loops and conditionals, including nested examples.
-
-The steps performed include:
-
-- Create a KFP pipeline:
-    - Use control flow components
-- Compile the KFP pipeline.
-- Execute the KFP pipeline using `Vertex AI Pipelines`
-
-[Metrics visualization and run comparison using the KFP SDK](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/pipelines/metrics_viz_run_compare_kfp.ipynb)
-
-Learn how to use the KFP SDK to build pipelines that generate evaluation metrics.
-
-The steps performed include:
-
-- Create KFP components:
-    - Generate ROC curve and confusion matrix visualizations for classification results
-    - Write metrics
-- Create KFP pipelines.
-- Execute KFP pipelines
-- Compare metrics across pipeline runs
-
-[Pipelines introduction for KFP](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/pipelines/pipelines_intro_kfp.ipynb)
-
-Learn how to use the KFP SDK to build pipelines that generate evaluation metrics.
-
-The steps performed include:
-
-- Define and compile a `Vertex AI` pipeline.
-- Specify which service account to use for a pipeline run.
-
-### Vertex AI Vizier 
-
-
-[Optimizing multiple objectives with Vertex AI Vizier](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/vizier/gapic-vizier-multi-objective-optimization.ipynb)
-
-Learn how to use `Vertex AI Vizier` to optimize a multi-objective study.
-
-
-
-### Vertex Explainable AI  Tabular data 
-
-
-[AutoML training tabular binary classification model for batch explanation](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/explainable_ai/sdk_automl_tabular_binary_classification_batch_explain.ipynb)
-
-Learn to use `AutoML` to create a tabular binary classification model from a Python script, and then learn to use `Vertex AI Batch Prediction` to make predictions with explanations.
-
-The steps performed include:
-
-- Create a `Vertex Dataset` resource.
-- Train an `AutoML` tabular binary classification model.
-- View the model evaluation metrics for the trained model.
-- Make a batch prediction request with explainability.
-
-
-* Prediction Service: Does an on-demand prediction for the entire set of instances (i.e., one or more data items) and returns the results in real-time.
-
-* Batch Prediction Service: Does a queued (batch) prediction for the entire set of instances in the background and stores the results in a Cloud Storage bucket when ready.
-
-[AutoML training tabular classification model for online explanation](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/explainable_ai/sdk_automl_tabular_classification_online_explain.ipynb)
-
-Learn how to use `AutoML` to create a tabular binary classification model from a Python script, and then learn to use `Vertex AI Online Prediction` to make online predictions with explanations.
-
-The steps performed include:
-
-- Create a `Vertex Dataset` resource.
-- Train an `AutoML` tabular binary classification model.
-- View the model evaluation metrics for the trained model.
-- Create a serving `Endpoint` resource.
-- Deploy the `Model` resource to a serving `Endpoint` resource.
-- Make an online prediction request with explainability.
-- Undeploy the `Model` resource.
-
-### Vertex Explainable AI  Image data 
-
-
-[Custom training image classification model for batch prediction with explainabilty](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/explainable_ai/sdk_custom_image_classification_batch_explain.ipynb)
-
-Learn to use `Vertex AI Training and Explainable AI` to create a custom image classification model with explanations, and then you learn to use `Vertex AI Batch Prediction` to make a batch prediction request with explanations.
-
-The steps performed include:
-
-- Create a `Vertex AI` custom job for training a TensorFlow model.
-- View the model evaluation for the trained model.
-- Set explanation parameters for when the model is deployed.
-- Upload the trained model artifacts and explanation parameters as a `Model` resource.
-- Make a batch prediction with explanations.
-
-[Custom training image classification model for online prediction with explainabilty](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/explainable_ai/sdk_custom_image_classification_online_explain.ipynb)
-
-Learn how to use `Vertex AI Training and Explainable AI` to create a custom image classification model with explanations, and then you learn to use `Vertex AI Prediction` to make an online prediction request with explanations.
-
-The steps performed include:
-
-- Create a `Vertex AI` custom job for training a TensorFlow model.
-- View the model evaluation for the trained model.
-- Set explanation parameters for when the model is deployed.
-- Upload the trained model artifacts and explanations as a `Model` resource.
-- Create a serving `Endpoint` resource.
-- Deploy the `Model` resource to a serving `Endpoint` resource.
-- Make a prediction with explanation.
-- Undeploy the `Model` resource.
-
-### Vertex Explainable AI  Tabular data 
-
-
-[Custom training tabular regression model for batch prediction with explainabilty](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/explainable_ai/sdk_custom_tabular_regression_batch_explain.ipynb)
-
-Learn how to use `Vertex AI Training and Explainable AI` to create a custom image classification model with explanations, and then you learn to use `Vertex AI Batch Prediction` to make a batch prediction request with explanations.
-
-The steps performed include:
-
-- Create a `Vertex AI` custom job for training a TensorFlow model.
-- View the model evaluation for the trained model.
-- Set explanation parameters for when the model is deployed.
-- Upload the trained model artifacts and explanations as a `Model` resource.
-- Make a batch prediction with explanations.
-
-### Vertex ML Metadata 
-
-
-[Track parameters and metrics for custom training jobs](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/ml_metadata/sdk-metric-parameter-tracking-for-custom-jobs.ipynb)
-
-Learn how to use Vertex AI SDK for Python to:
-
-The steps performed include:
-- Track training parameters and prediction metrics for a custom training job.
-- Extract and perform analysis for all parameters and metrics within an Experiment.
-
-
-
-
- 

notebooks/official/automl/README.md

@@ -1,4 +1,5 @@
-[AutoML Tabular Training and Prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl-tabular-classification.ipynb)
+
+[AutoML Tabular training and prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl-tabular-classification.ipynb)
 
 ```
 Learn how to train and make predictions on an AutoML model based on a tabular dataset.
@@ -13,6 +14,8 @@ The steps performed include the following:
 
 ```
 
+   Learn more about [Classification for tabular data](https://cloud.google.com/vertex-ai/docs/tabular-data/classification-regression/overview).
+
 
 [Create, train, and deploy an AutoML text classification model](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl-text-classification.ipynb)
 
@@ -31,11 +34,86 @@ The steps performed include:
 
 ```
 
+   Learn more about [Classification for text data](https://cloud.google.com/vertex-ai/docs/training-overview#classification_for_text).
 
-[AutoML training video classification model for batch prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/sdk_automl_video_classification_batch.ipynb)
+
+[Compare Vertex AI Forecasting and BigQuery ML ARIMA_PLUS](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_forecasting_bqml_arima_plus_comparison.ipynb)
 
 ```
-Learn how to create an AutoML video classification model from a Python script, and then do a batch prediction using the Vertex AI SDK.
+Learn how to create an BigQuery ML ARIMA_PLUS model using a training Vertex AI Pipeline from Google Cloud Pipeline Components , and then do a batch prediction using the corresponding prediction pipeline.
+
+The steps performed are:
+
+- Train the BigQuery ML ARIMA_PLUS model.
+- View BigQuery ML model evaluation.
+- Make a batch prediction with the BigQuery ML model.
+- Create a Vertex AI `Dataset` resource.
+- Train the Vertex AI Forecasting model.
+- View the Model evaluation.
+- Make a batch prediction with the Model.
+
+```
+
+   Learn more about [BQML ARIMA+ forecasting for tabular data](https://cloud.google.com/vertex-ai/docs/tabular-data/forecasting-arima/overview).
+
+
+[AutoML Tabular Workflow pipelines](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_tabular_on_vertex_pipelines.ipynb)
+
+```
+Learn how to create two regression models using Vertex AI Pipelines downloaded from Google Cloud Pipeline Components .
+
+The steps performed are:
+
+- Create a training pipeline that reduces the search space from the default to save time.
+- Create a training pipeline that reuses the architecture search results from the previous pipeline to save time.
+
+```
+
+   Learn more about [Tabular Workflow for E2E AutoML](https://cloud.google.com/vertex-ai/docs/tabular-data/tabular-workflows/e2e-automl).
+
+
+[Get started with AutoML Training](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/get_started_automl_training.ipynb)
+
+```
+Learn how to use `AutoML` for training with `Vertex AI`.
+
+The steps performed include:
+
+- Train an image model
+- Export the image model as an edge model
+- Train a tabular model
+- Export the tabular model as a cloud model
+- Train a text model
+- Train a video model
+
+```
+
+   Learn more about [AutoML training](https://cloud.google.com/vertex-ai/docs/training-overview).
+
+
+[AutoML training hierarchical forecasting for batch prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/sdk_automl_forecasting_hierarchical_batch.ipynb)
+
+```
+In this tutorial, you create an AutoML hierarchical forecasting model and deploy it for batch prediction using the Vertex AI SDK for Python.
+
+The steps performed include:
+
+- Create a Vertex AI `TimeSeriesDataset` resource.
+- Train the model.
+- View the model evaluation.
+- Deploy the `Model` resource to a serving `Endpoint` resource.
+- Make a prediction.
+- Undeploy the `Model`.
+
+```
+
+   Learn more about [Hierarchical forecasting for tabular data](https://cloud.google.com/vertex-ai/docs/tabular-data/forecasting/hierarchical).
+
+
+[AutoML training image object detection model for batch prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/sdk_automl_image_object_detection_batch.ipynb)
+
+```
+In this tutorial, you create an AutoML image object detection model from a Python script, and then do a batch prediction using the Vertex AI SDK.
 
 The steps performed include:
 
@@ -46,22 +124,7 @@ The steps performed include:
 
 ```
 
-
-[AutoML training text entity extraction model for online prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/sdk_automl_text_entity_extraction_online.ipynb)
-
-```
-Learn how to create an AutoML text entity extraction model and deploy for online prediction from a Python script using the Vertex SDK.
-
-The steps performed include:
-
-- Create a Vertex `Dataset` resource.
-- Train the model.
-- View the model evaluation.
-- Deploy the `Model` resource to a serving `Endpoint` resource.
-- Make a prediction.
-- Undeploy the `Model`.
-
-```
+   Learn more about [Object detection for image data](https://cloud.google.com/vertex-ai/docs/training-overview#object_detection_for_images).
 
 
 [AutoML tabular forecasting model for batch prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/sdk_automl_tabular_forecasting_batch.ipynb)
@@ -78,32 +141,64 @@ The steps performed include:
 
 ```
 
+   Learn more about [Forecasting for tabular data](https://cloud.google.com/vertex-ai/docs/tabular-data/forecasting/overview).
 
-[AutoML training video action recognition model for batch prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/sdk_automl_video_action_recognition_batch.ipynb)
+
+[AutoML training tabular regression model for batch prediction using BigQuery](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/sdk_automl_tabular_regression_batch_bq.ipynb)
 
 ```
-Learn how to create an AutoML video action recognition model from a Python script, and then do a batch prediction using the Vertex AI SDK.
+Learn how to create an AutoML tabular regression model and deploy it for batch prediction using the Vertex AI SDK for Python.
 
 The steps performed include:
 
-- Create a `Vertex AI Dataset` resource.
+- Create a Vertex AI `Dataset` resource.
 - Train the model.
 - View the model evaluation.
-- Make a batch prediction.
-```
-
-
-[AutoML Tabular Pipeline](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_tabular_on_vertex_pipelines.ipynb)
+- Deploy the `Model` resource to a serving `Endpoint` resource.
+- Make a prediction.
+- Undeploy the `Model`.
 
 ```
-Learn how to create two regression models using [Vertex Pipelines](https://cloud.
 
-The steps performed are:
+   Learn more about [Regression for tabular data](https://cloud.google.com/vertex-ai/docs/tabular-data/classification-regression/overview).
 
-- Create a training pipeline that reduces the search space from the default to save time.
-- Create a training pipeline that reuses the architecture search results from the previous pipeline to save time.
+
+[AutoML training tabular regression model for online prediction using BigQuery](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/sdk_automl_tabular_regression_online_bq.ipynb)
 
 ```
+Learn how to create an AutoML tabular regression model and deploy for online prediction from a Python script using the Vertex AI SDK.
+
+The steps performed include:
+
+- Create a Vertex `Dataset` resource.
+- Train the model.
+- View the model evaluation.
+- Deploy the `Model` resource to a serving `Endpoint` resource.
+- Make a prediction.
+- Undeploy the `Model`.
+
+```
+
+   Learn more about [Regression for tabular data](https://cloud.google.com/vertex-ai/docs/tabular-data/classification-regression/overview).
+
+
+[AutoML training text entity extraction model for online prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/sdk_automl_text_entity_extraction_online.ipynb)
+
+```
+Learn how to create an AutoML text entity extraction model and deploy for online prediction from a Python script using the Vertex AI SDK.
+
+The steps performed include:
+
+- Create a Vertex `Dataset` resource.
+- Train the model.
+- View the model evaluation.
+- Deploy the `Model` resource to a serving `Endpoint` resource.
+- Make a prediction.
+- Undeploy the `Model`.
+
+```
+
+   Learn more about [Entity extraction for text data](https://cloud.google.com/vertex-ai/docs/training-overview#entity_extraction_for_text).
 
 
 [Training an AutoML text sentiment analysis model for online predictions](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/sdk_automl_text_sentiment_analysis_online.ipynb)
@@ -123,41 +218,42 @@ The steps performed include:
 
 ```
 
+   Learn more about [Sentiment analysis for text data](https://cloud.google.com/vertex-ai/docs/training-overview#sentiment_analysis_for_text).
 
-[Compare Vertex AI Forecasting and BigQuery ML ARIMA_PLUS](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_forecasting_bqml_arima_plus_comparison.ipynb)
+
+[AutoML training video action recognition model for batch prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/sdk_automl_video_action_recognition_batch.ipynb)
 
 ```
-Learn how to create an BigQuery ML ARIMA_PLUS model using a training [Vertex AI Pipeline](https://cloud.
+Learn how to create an AutoML video action recognition model from a Python script, and then do a batch prediction using the Vertex AI SDK.
 
-The steps performed are:
+The steps performed include:
 
-- Train the BigQuery ML ARIMA_PLUS model.
-- View BigQuery ML model evaluation.
-- Make a batch prediction with the BigQuery ML model.
-- Create a Vertex AI `Dataset` resource.
-- Train the Vertex AI Forecasting model.
-- View the Model evaluation.
-- Make a batch prediction with the Model.
+- Create a `Vertex AI Dataset` resource.
+- Train the model.
+- View the model evaluation.
+- Make a batch prediction.
 
 ```
 
+   Learn more about [Action recognition for video data](https://cloud.google.com/vertex-ai/docs/training-overview#action_recognition_for_videos).
 
-[AutoML training tabular regression model for online prediction using BigQuery](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/sdk_automl_tabular_regression_online_bq.ipynb)
+
+[AutoML training video classification model for batch prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/sdk_automl_video_classification_batch.ipynb)
 
 ```
-Learn how to create an AutoML tabular regression model and deploy for online prediction from a Python script using the Vertex AI SDK.
+Learn how to create an AutoML video classification model from a Python script, and then do a batch prediction using the Vertex AI SDK.
 
 The steps performed include:
 
 - Create a Vertex `Dataset` resource.
 - Train the model.
 - View the model evaluation.
-- Deploy the `Model` resource to a serving `Endpoint` resource.
-- Make a prediction.
-- Undeploy the `Model`.
+- Make a batch prediction.
 
 ```
 
+   Learn more about [Classification for video data](https://cloud.google.com/vertex-ai/docs/training-overview#classification_for_videos).
+
 
 [AutoML training video object tracking model for batch prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/sdk_automl_video_object_tracking_batch.ipynb)
 
@@ -170,21 +266,8 @@ The steps performed include:
 - Train the model.
 - View the model evaluation.
 - Make a batch prediction.
-```
-
-
-[AutoML training tabular regression model for batch prediction using BigQuery](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/sdk_automl_tabular_regression_batch_bq.ipynb)
 
 ```
-Learn how to create an AutoML tabular regression model and deploy it for batch prediction using the Vertex AI SDK for Python.
 
-The steps performed include:
+   Learn more about [Object tracking for video data](https://cloud.google.com/vertex-ai/docs/training-overview#object_tracking_for_videos).
 
-- Create a Vertex AI `Dataset` resource.
-- Train the model.
-- View the model evaluation.
-- Deploy the `Model` resource to a serving `Endpoint` resource.
-- Make a prediction.
-- Undeploy the `Model`.
-
-```

notebooks/official/automl/automl-tabular-classification.ipynb

@@ -63,7 +63,9 @@
         "\n",
         "This tutorial demonstrates how to use the Vertex AI Python client library to train and deploy a tabular classification model for online prediction.\n",
         "\n",
-        "**Note**: you may incur charges for training, prediction, storage, or usage of other Google Cloud products in connection with testing this SDK."
+        "**Note**: you may incur charges for training, prediction, storage, or usage of other Google Cloud products in connection with testing this SDK.\n",
+        "\n",
+        "Learn more about [Classification for tabular data](https://cloud.google.com/vertex-ai/docs/tabular-data/classification-regression/overview)."
       ]
     },
     {

notebooks/official/automl/automl-text-classification.ipynb

@@ -68,7 +68,9 @@
       "source": [
         "## Overview\n",
         "\n",
-        "This notebook walks you through the major phases of building and using an AutoML text classification model on [Vertex AI](https://cloud.google.com/vertex-ai/docs/). \n"
+        "This notebook walks you through the major phases of building and using an AutoML text classification model on [Vertex AI](https://cloud.google.com/vertex-ai/docs/). \n",
+        "\n",
+        "Learn more about [Classification for text data](https://cloud.google.com/vertex-ai/docs/training-overview#classification_for_text)."
       ]
     },
     {

notebooks/official/automl/automl_forecasting_bqml_arima_plus_comparison.ipynb

@@ -61,7 +61,9 @@
       "source": [
         "## Overview\n",
         "\n",
-        "In this tutorial, you take on the role of a store planner who must determine how much inventory they will need to order for each of their products and stores for November 2019. You accomplish this by training forecasting models using historical sales data. You start with a baseline model using BigQuery ML (BQML) [ARIMA_PLUS](https://cloud.google.com/bigquery-ml/docs/reference/standard-sql/bigqueryml-syntax-create-time-series) and then compare it against a [Vertex AI Forecasting](https://cloud.google.com/vertex-ai/docs/tabular-data/forecasting/overview) model."
+        "In this tutorial, you take on the role of a store planner who must determine how much inventory they will need to order for each of their products and stores for November 2019. You accomplish this by training forecasting models using historical sales data. You start with a baseline model using BigQuery ML (BQML) [ARIMA_PLUS](https://cloud.google.com/bigquery-ml/docs/reference/standard-sql/bigqueryml-syntax-create-time-series) and then compare it against a [Vertex AI Forecasting](https://cloud.google.com/vertex-ai/docs/tabular-data/forecasting/overview) model.\n",
+        "\n",
+        "Learn more about [BQML ARIMA+ forecasting for tabular data](https://cloud.google.com/vertex-ai/docs/tabular-data/forecasting-arima/overview)."
       ]
     },
     {

notebooks/official/automl/automl_image_classification_batch_prediction.ipynb

@@ -0,0 +1,884 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "copyright"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2021 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "title"
+      },
+      "source": [
+        "# AutoML training image classification model for batch prediction\n",
+        "\n",
+        "<table align=\"left\">\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_image_classification_batch_prediction.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_image_classification_batch_prediction.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/workbench/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/main/notebooks/official/automl/automl_image_classification_online_prediction.ipynb\">\n",
+        "       <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "      Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>\n",
+        "<br/><br/><br/>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "overview:automl"
+      },
+      "source": [
+        "## Overview\n",
+        "\n",
+        "\n",
+        "This tutorial demonstrates how to use the Vertex AI SDK to create image classification models and do batch prediction using a Google Cloud [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) model.\n",
+        "\n",
+        "Learn more about [Get predictions from an image classification model](https://cloud.google.com/vertex-ai/docs/image-data/classification/get-predictions)."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "objective:automl,training,batch_prediction"
+      },
+      "source": [
+        "### Objective\n",
+        "\n",
+        "In this tutorial, you create an AutoML image classification model from a Python script, and then do a batch prediction using the Vertex SDK. You can alternatively create and deploy models using the `gcloud` command-line tool or online using the Cloud Console.\n",
+        "\n",
+        "The steps performed include:\n",
+        "\n",
+        "- Create a Vertex `Dataset` resource.\n",
+        "- Train the model.\n",
+        "- View the model evaluation.\n",
+        "- Make a batch prediction.\n",
+        "\n",
+        "There is one key difference between using batch prediction and using online prediction:\n",
+        "\n",
+        "* Prediction Service: Does an on-demand prediction for the entire set of instances (i.e., one or more data items) and returns the results in real-time.\n",
+        "\n",
+        "* Batch Prediction Service: Does a queued (batch) prediction for the entire set of instances in the background and stores the results in a Cloud Storage bucket when ready."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "dataset:flowers,icn"
+      },
+      "source": [
+        "### Dataset\n",
+        "\n",
+        "The dataset used for this tutorial is the [Flowers dataset](https://www.tensorflow.org/datasets/catalog/tf_flowers) from [TensorFlow Datasets](https://www.tensorflow.org/datasets/catalog/overview). The version of the dataset you will use in this tutorial is stored in a public Cloud Storage bucket. The trained model predicts the type of flower an image is from a class of five flowers: daisy, dandelion, rose, sunflower, or tulip."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "costs"
+      },
+      "source": [
+        "### Costs\n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI\n",
+        "* Cloud Storage\n",
+        "\n",
+        "Learn about [Vertex AI\n",
+        "pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage\n",
+        "pricing](https://cloud.google.com/storage/pricing), and use the [Pricing\n",
+        "Calculator](https://cloud.google.com/products/calculator/)\n",
+        "to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "install_aip:mbsdk"
+      },
+      "source": [
+        "## Installation\n",
+        "\n",
+        "Install the latest version of Vertex AI SDK for Python."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "install_aip:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "import os\n",
+        "\n",
+        "USER_FLAG = \"\"\n",
+        "\n",
+        "! pip3 install --upgrade google-cloud-aiplatform $USER_FLAG\n",
+        "\n",
+        "if os.environ[\"IS_TESTING\"]:\n",
+        "    ! pip3 install --upgrade tensorflow $USER_FLAG"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "restart"
+      },
+      "source": [
+        "### Colab only: Uncomment the following cell to restart the kernel"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "D-ZBOjErv5mM"
+      },
+      "outputs": [],
+      "source": [
+        "# Automatically restart kernel after installs so that your environment can access the new packages\n",
+        "# import IPython\n",
+        "\n",
+        "# app = IPython.Application.instance()\n",
+        "# app.kernel.do_shutdown(True)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "before_you_begin:nogpu"
+      },
+      "source": [
+        "## Before you begin"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "before_you_begin:nogpu"
+      },
+      "source": [
+        "### Set your project ID\n",
+        "\n",
+        "**If you don't know your project ID**, try the following:\n",
+        "* Run `gcloud config list`.\n",
+        "* Run `gcloud projects list`.\n",
+        "* See the support page: [Locate the project ID](https://support.google.com/googleapi/answer/7014113)"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "set_project_id"
+      },
+      "outputs": [],
+      "source": [
+        "PROJECT_ID = \"[your-project-id]\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# Set the project id\n",
+        "! gcloud config set project {PROJECT_ID}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "region"
+      },
+      "source": [
+        "#### Region\n",
+        "\n",
+        "You can also change the `REGION` variable used by Vertex AI. Learn more about [Vertex AI regions](https://cloud.google.com/vertex-ai/docs/general/locations)."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "region"
+      },
+      "outputs": [],
+      "source": [
+        "REGION = \"us-central1\"  # @param {type: \"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "gcp_authenticate"
+      },
+      "source": [
+        "### Authenticate your Google Cloud account\n",
+        "\n",
+        "Depending on your Jupyter environment, you may have to manually authenticate. Follow the relevant instructions below."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "FvQeFm3Gv5mR"
+      },
+      "source": [
+        "**1. Vertex AI Workbench**\n",
+        "* Do nothing as you are already authenticated."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "ad1138a125ea"
+      },
+      "source": [
+        "**2. Local JupyterLab instance, uncomment and run:**"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "ce6043da7b33"
+      },
+      "outputs": [],
+      "source": [
+        "# ! gcloud auth login"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0367eac06a10"
+      },
+      "source": [
+        "**3. Colab, uncomment and run:**"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "21ad4dbb4a61"
+      },
+      "outputs": [],
+      "source": [
+        "# from google.colab import auth\n",
+        "# auth.authenticate_user()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "c13224697bfb"
+      },
+      "source": [
+        "**4. Service account or other**\n",
+        "* See how to grant Cloud Storage permissions to your service account at https://cloud.google.com/storage/docs/gsutil/commands/iam#ch-examples."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "bucket:mbsdk"
+      },
+      "source": [
+        "### Create a Cloud Storage bucket\n",
+        "\n",
+        "Create a storage bucket to store intermediate artifacts such as datasets."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "bucket"
+      },
+      "outputs": [],
+      "source": [
+        "BUCKET_URI = f\"gs://your-bucket-name-unique-{PROJECT_ID}\"  # @param {type:\"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "create_bucket"
+      },
+      "source": [
+        "**Only if your bucket doesn't already exist**: Run the following cell to create your Cloud Storage bucket."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "create_bucket"
+      },
+      "outputs": [],
+      "source": [
+        "! gsutil mb -l $REGION $BUCKET_URI"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "setup_vars"
+      },
+      "source": [
+        "### Set up variables\n",
+        "\n",
+        "Next, set up some variables used throughout the tutorial.\n",
+        "### Import libraries and define constants"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "import_aip:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "import google.cloud.aiplatform as aiplatform"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "init_aip:mbsdk"
+      },
+      "source": [
+        "## Initialize Vertex AI SDK for Python\n",
+        "\n",
+        "Initialize the Vertex AI SDK for Python for your project and corresponding bucket."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "init_aip:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "aiplatform.init(project=PROJECT_ID, staging_bucket=BUCKET_URI)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "tutorial_start:automl"
+      },
+      "source": [
+        "# Tutorial\n",
+        "\n",
+        "Now you are ready to start creating your own AutoML image classification model."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "import_file:u_dataset,csv"
+      },
+      "source": [
+        "#### Location of Cloud Storage training data.\n",
+        "\n",
+        "Now set the variable `IMPORT_FILE` to the location of the CSV index file in Cloud Storage."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "import_file:flowers,csv,icn"
+      },
+      "outputs": [],
+      "source": [
+        "IMPORT_FILE = (\n",
+        "    \"gs://cloud-samples-data/vision/automl_classification/flowers/all_data_v2.csv\"\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "quick_peek:csv"
+      },
+      "source": [
+        "#### Quick peek at your data\n",
+        "\n",
+        "This tutorial uses a version of the Flowers dataset that is stored in a public Cloud Storage bucket, using a CSV index file.\n",
+        "\n",
+        "Start by doing a quick peek at the data. You count the number of examples by counting the number of rows in the CSV index file  (`wc -l`) and then peek at the first few rows."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "quick_peek:csv"
+      },
+      "outputs": [],
+      "source": [
+        "if \"IMPORT_FILES\" in globals():\n",
+        "    FILE = IMPORT_FILES[0]\n",
+        "else:\n",
+        "    FILE = IMPORT_FILE\n",
+        "\n",
+        "count = ! gsutil cat $FILE | wc -l\n",
+        "print(\"Number of Examples\", int(count[0]))\n",
+        "\n",
+        "print(\"First 10 rows\")\n",
+        "! gsutil cat $FILE | head"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "create_dataset:image,icn"
+      },
+      "source": [
+        "### Create the Dataset\n",
+        "\n",
+        "Next, create the `Dataset` resource using the `create` method for the `ImageDataset` class, which takes the following parameters:\n",
+        "\n",
+        "- `display_name`: The human readable name for the `Dataset` resource.\n",
+        "- `gcs_source`: A list of one or more dataset index files to import the data items into the `Dataset` resource.\n",
+        "- `import_schema_uri`: The data labeling schema for the data items.\n",
+        "\n",
+        "This operation may take several minutes."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "create_dataset:image,icn"
+      },
+      "outputs": [],
+      "source": [
+        "dataset = aiplatform.ImageDataset.create(\n",
+        "    display_name=\"Flowers\",\n",
+        "    gcs_source=[IMPORT_FILE],\n",
+        "    import_schema_uri=aiplatform.schema.dataset.ioformat.image.single_label_classification,\n",
+        ")\n",
+        "\n",
+        "print(dataset.resource_name)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "create_automl_pipeline:image,icn"
+      },
+      "source": [
+        "### Create and run training pipeline\n",
+        "\n",
+        "To train an AutoML model, you perform two steps: 1) create a training pipeline, and 2) run the pipeline.\n",
+        "\n",
+        "#### Create training pipeline\n",
+        "\n",
+        "An AutoML training pipeline is created with the `AutoMLImageTrainingJob` class, with the following parameters:\n",
+        "\n",
+        "- `display_name`: The human readable name for the `TrainingJob` resource.\n",
+        "- `prediction_type`: The type task to train the model for.\n",
+        "  - `classification`: An image classification model.\n",
+        "  - `object_detection`: An image object detection model.\n",
+        "- `multi_label`: If a classification task, whether single (`False`) or multi-labeled (`True`).\n",
+        "- `model_type`: The type of model for deployment.\n",
+        "  - `CLOUD`: Deployment on Google Cloud\n",
+        "  - `CLOUD_HIGH_ACCURACY_1`: Optimized for accuracy over latency for deployment on Google Cloud.\n",
+        "  - `CLOUD_LOW_LATENCY_`: Optimized for latency over accuracy for deployment on Google Cloud.\n",
+        "  - `MOBILE_TF_VERSATILE_1`: Deployment on an edge device.\n",
+        "  - `MOBILE_TF_HIGH_ACCURACY_1`:Optimized for accuracy over latency for deployment on an edge device.\n",
+        "  - `MOBILE_TF_LOW_LATENCY_1`: Optimized for latency over accuracy for deployment on an edge device.\n",
+        "- `base_model`: (optional) Transfer learning from existing `Model` resource -- supported for image classification only.\n",
+        "\n",
+        "The instantiated object is the DAG (directed acyclic graph) for the training job."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "create_automl_pipeline:image,icn"
+      },
+      "outputs": [],
+      "source": [
+        "dag = aiplatform.AutoMLImageTrainingJob(\n",
+        "    display_name=\"flowers\",\n",
+        "    prediction_type=\"classification\",\n",
+        "    multi_label=False,\n",
+        "    model_type=\"CLOUD\",\n",
+        "    base_model=None,\n",
+        ")\n",
+        "\n",
+        "print(dag)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "run_automl_pipeline:image"
+      },
+      "source": [
+        "#### Run the training pipeline\n",
+        "\n",
+        "Next, you run the DAG to start the training job by invoking the method `run`, with the following parameters:\n",
+        "\n",
+        "- `dataset`: The `Dataset` resource to train the model.\n",
+        "- `model_display_name`: The human readable name for the trained model.\n",
+        "- `training_fraction_split`: The percentage of the dataset to use for training.\n",
+        "- `test_fraction_split`: The percentage of the dataset to use for test (holdout data).\n",
+        "- `validation_fraction_split`: The percentage of the dataset to use for validation.\n",
+        "- `budget_milli_node_hours`: (optional) Maximum training time specified in unit of millihours (1000 = hour).\n",
+        "- `disable_early_stopping`: If `True`, training maybe completed before using the entire budget if the service believes it cannot further improve on the model objective measurements.\n",
+        "\n",
+        "The `run` method when completed returns the `Model` resource.\n",
+        "\n",
+        "The execution of the training pipeline will take upto 20 minutes."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "run_automl_pipeline:image"
+      },
+      "outputs": [],
+      "source": [
+        "model = dag.run(\n",
+        "    dataset=dataset,\n",
+        "    model_display_name=\"flowers\",\n",
+        "    training_fraction_split=0.8,\n",
+        "    validation_fraction_split=0.1,\n",
+        "    test_fraction_split=0.1,\n",
+        "    budget_milli_node_hours=8000,\n",
+        "    disable_early_stopping=False,\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "evaluate_the_model:mbsdk"
+      },
+      "source": [
+        "## Review model evaluation scores\n",
+        "\n",
+        "After your model training has finished, you can review the evaluation scores for it using the `list_model_evaluations()` method. This method will return an iterator for each evaluation slice."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "make_prediction"
+      },
+      "source": [
+        "## Send a batch prediction request\n",
+        "\n",
+        "Send a batch prediction to your deployed model."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "evaluate_the_model:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "model_evaluations = model.list_model_evaluations()\n",
+        "\n",
+        "for model_evaluation in model_evaluations:\n",
+        "    print(model_evaluation.to_dict())"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "get_test_items:batch_prediction"
+      },
+      "source": [
+        "### Get test item(s)\n",
+        "\n",
+        "Now do a batch prediction to your Vertex model. You will use arbitrary examples out of the dataset as a test items. Don't be concerned that the examples were likely used in training the model -- we just want to demonstrate how to make a prediction."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "get_test_items:automl,icn,csv"
+      },
+      "outputs": [],
+      "source": [
+        "test_items = !gsutil cat $IMPORT_FILE | head -n2\n",
+        "if len(str(test_items[0]).split(\",\")) == 3:\n",
+        "    _, test_item_1, test_label_1 = str(test_items[0]).split(\",\")\n",
+        "    _, test_item_2, test_label_2 = str(test_items[1]).split(\",\")\n",
+        "else:\n",
+        "    test_item_1, test_label_1 = str(test_items[0]).split(\",\")\n",
+        "    test_item_2, test_label_2 = str(test_items[1]).split(\",\")\n",
+        "\n",
+        "print(test_item_1, test_label_1)\n",
+        "print(test_item_2, test_label_2)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "copy_test_items:batch_prediction"
+      },
+      "source": [
+        "### Copy test item(s)\n",
+        "\n",
+        "For the batch prediction, copy the test items over to your Cloud Storage bucket."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "copy_test_items:batch_prediction"
+      },
+      "outputs": [],
+      "source": [
+        "file_1 = test_item_1.split(\"/\")[-1]\n",
+        "file_2 = test_item_2.split(\"/\")[-1]\n",
+        "\n",
+        "! gsutil cp $test_item_1 $BUCKET_URI/$file_1\n",
+        "! gsutil cp $test_item_2 $BUCKET_URI/$file_2\n",
+        "\n",
+        "test_item_1 = BUCKET_URI + \"/\" + file_1\n",
+        "test_item_2 = BUCKET_URI + \"/\" + file_2"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "make_batch_file:automl,image"
+      },
+      "source": [
+        "### Make the batch input file\n",
+        "\n",
+        "Now make a batch input file, which you will store in your local Cloud Storage bucket. The batch input file can be either CSV or JSONL. You will use JSONL in this tutorial. For JSONL file, you make one dictionary entry per line for each data item (instance). The dictionary contains the key/value pairs:\n",
+        "\n",
+        "- `content`: The Cloud Storage path to the image.\n",
+        "- `mime_type`: The content type. In our example, it is a `jpeg` file.\n",
+        "\n",
+        "For example:\n",
+        "\n",
+        "                        {'content': '[your-bucket]/file1.jpg', 'mime_type': 'jpeg'}"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "make_batch_file:automl,image"
+      },
+      "outputs": [],
+      "source": [
+        "import json\n",
+        "\n",
+        "import tensorflow as tf\n",
+        "\n",
+        "gcs_input_uri = BUCKET_URI + \"/test.jsonl\"\n",
+        "with tf.io.gfile.GFile(gcs_input_uri, \"w\") as f:\n",
+        "    data = {\"content\": test_item_1, \"mime_type\": \"image/jpeg\"}\n",
+        "    f.write(json.dumps(data) + \"\\n\")\n",
+        "    data = {\"content\": test_item_2, \"mime_type\": \"image/jpeg\"}\n",
+        "    f.write(json.dumps(data) + \"\\n\")\n",
+        "\n",
+        "print(gcs_input_uri)\n",
+        "! gsutil cat $gcs_input_uri"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "batch_request:mbsdk"
+      },
+      "source": [
+        "### Make the batch prediction request\n",
+        "\n",
+        "Now that your Model resource is trained, you can make a batch prediction by invoking the batch_predict() method, with the following parameters:\n",
+        "\n",
+        "- `job_display_name`: The human readable name for the batch prediction job.\n",
+        "- `gcs_source`: A list of one or more batch request input files.\n",
+        "- `gcs_destination_prefix`: The Cloud Storage location for storing the batch prediction resuls.\n",
+        "- `sync`: If set to True, the call will block while waiting for the asynchronous batch job to complete."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "batch_request:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "batch_predict_job = model.batch_predict(\n",
+        "    job_display_name=\"flowers\",\n",
+        "    gcs_source=gcs_input_uri,\n",
+        "    gcs_destination_prefix=BUCKET_URI,\n",
+        "    sync=False,\n",
+        ")\n",
+        "\n",
+        "print(batch_predict_job)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "batch_request_wait:mbsdk"
+      },
+      "source": [
+        "### Wait for completion of batch prediction job\n",
+        "\n",
+        "Next, wait for the batch job to complete. Alternatively, one can set the parameter `sync` to `True` in the `batch_predict()` method to block until the batch prediction job is completed."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "batch_request_wait:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "batch_predict_job.wait()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "get_batch_prediction:mbsdk,icn"
+      },
+      "source": [
+        "### Get the predictions\n",
+        "\n",
+        "Next, get the results from the completed batch prediction job.\n",
+        "\n",
+        "The results are written to the Cloud Storage output bucket you specified in the batch prediction request. You call the method iter_outputs() to get a list of each Cloud Storage file generated with the results. Each file contains one or more prediction requests in a JSON format:\n",
+        "\n",
+        "- `content`: The prediction request.\n",
+        "- `prediction`: The prediction response.\n",
+        " - `ids`: The internal assigned unique identifiers for each prediction request.\n",
+        " - `displayNames`: The class names for each class label.\n",
+        " - `confidences`: The predicted confidence, between 0 and 1, per class label."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "get_batch_prediction:mbsdk,icn"
+      },
+      "outputs": [],
+      "source": [
+        "import json\n",
+        "\n",
+        "import tensorflow as tf\n",
+        "\n",
+        "bp_iter_outputs = batch_predict_job.iter_outputs()\n",
+        "\n",
+        "prediction_results = list()\n",
+        "for blob in bp_iter_outputs:\n",
+        "    if blob.name.split(\"/\")[-1].startswith(\"prediction\"):\n",
+        "        prediction_results.append(blob.name)\n",
+        "\n",
+        "tags = list()\n",
+        "for prediction_result in prediction_results:\n",
+        "    gfile_name = f\"gs://{bp_iter_outputs.bucket.name}/{prediction_result}\"\n",
+        "    with tf.io.gfile.GFile(name=gfile_name, mode=\"r\") as gfile:\n",
+        "        for line in gfile.readlines():\n",
+        "            line = json.loads(line)\n",
+        "            print(line)\n",
+        "            break"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "cleanup:mbsdk"
+      },
+      "source": [
+        "# Cleaning up\n",
+        "\n",
+        "To clean up all Google Cloud resources used in this project, you can [delete the Google Cloud\n",
+        "project](https://cloud.google.com/resource-manager/docs/creating-managing-projects#shutting_down_projects) you used for the tutorial.\n",
+        "\n",
+        "Otherwise, you can delete the individual resources you created in this tutorial."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "cleanup:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "delete_bucket = False\n",
+        "\n",
+        "# Delete the dataset using the Vertex dataset object\n",
+        "dataset.delete()\n",
+        "\n",
+        "# Delete the model using the Vertex model object\n",
+        "model.delete()\n",
+        "\n",
+        "# Delete the AutoML trainig job\n",
+        "dag.delete()\n",
+        "\n",
+        "# Delete the batch prediction job\n",
+        "batch_predict_job.delete()\n",
+        "\n",
+        "if delete_bucket or os.getenv(\"IS_TESTING\"):\n",
+        "    ! gsutil rm -r $BUCKET_URI"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "name": "automl_image_classification_batch_prediction.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/official/automl/automl_image_classification_online_online_prediction.ipynb

@@ -0,0 +1,801 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "copyright"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2020 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "title"
+      },
+      "source": [
+        "# AutoML training image classification model for online prediction\n",
+        "\n",
+        "<table align=\"left\">\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_image_classification_online_prediction.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_image_classification_online_prediction.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/workbench/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/main/notebooks/official/automl//automl_image_classification_online_prediction.ipynb\">\n",
+        "       <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "      Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>\n",
+        "<br/><br/><br/>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "overview:automl"
+      },
+      "source": [
+        "## Overview\n",
+        "\n",
+        "\n",
+        "This tutorial demonstrates how to use the Vertex AI SDK to create image classification models and do online prediction using a Google Cloud [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) model.\n",
+        "\n",
+        "Learn more about [Get predictions from an image classification model](https://cloud.google.com/vertex-ai/docs/image-data/classification/get-predictions)."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "objective:automl,training,online_prediction"
+      },
+      "source": [
+        "### Objective\n",
+        "\n",
+        "In this tutorial, you create an AutoML image classification model and deploy for online prediction from a Python script using the Vertex SDK. You can alternatively create and deploy models using the `gcloud` command-line tool or online using the Cloud Console.\n",
+        "\n",
+        "The steps performed include:\n",
+        "\n",
+        "- Create a Vertex `Dataset` resource.\n",
+        "- Train the model.\n",
+        "- View the model evaluation.\n",
+        "- Deploy the `Model` resource to a serving `Endpoint` resource.\n",
+        "- Make a prediction.\n",
+        "- Undeploy the `Model`."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "dataset:flowers,icn"
+      },
+      "source": [
+        "### Dataset\n",
+        "\n",
+        "The dataset used for this tutorial is the [Flowers dataset](https://www.tensorflow.org/datasets/catalog/tf_flowers) from [TensorFlow Datasets](https://www.tensorflow.org/datasets/catalog/overview). The version of the dataset you will use in this tutorial is stored in a public Cloud Storage bucket. The trained model predicts the type of flower an image is from a class of five flowers: daisy, dandelion, rose, sunflower, or tulip."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "costs"
+      },
+      "source": [
+        "### Costs\n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI\n",
+        "* Cloud Storage\n",
+        "\n",
+        "Learn about [Vertex AI\n",
+        "pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage\n",
+        "pricing](https://cloud.google.com/storage/pricing), and use the [Pricing\n",
+        "Calculator](https://cloud.google.com/products/calculator/)\n",
+        "to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "install_aip:mbsdk"
+      },
+      "source": [
+        "## Installation\n",
+        "\n",
+        "Install the latest version of Vertex AI SDK for Python."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "install_aip:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "import os\n",
+        "\n",
+        "USER_FLAG = \"\"\n",
+        "\n",
+        "! pip3 install --upgrade google-cloud-aiplatform \\\n",
+        "                         tensorflow $USER_FLAG"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "restart"
+      },
+      "source": [
+        "### Colab only: Uncomment the following cell to restart the kernel"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "D-ZBOjErv5mM"
+      },
+      "outputs": [],
+      "source": [
+        "# Automatically restart kernel after installs so that your environment can access the new packages\n",
+        "# import IPython\n",
+        "\n",
+        "# app = IPython.Application.instance()\n",
+        "# app.kernel.do_shutdown(True)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "before_you_begin:nogpu"
+      },
+      "source": [
+        "## Before you begin"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "before_you_begin:nogpu"
+      },
+      "source": [
+        "### Set your project ID\n",
+        "\n",
+        "**If you don't know your project ID**, try the following:\n",
+        "* Run `gcloud config list`.\n",
+        "* Run `gcloud projects list`.\n",
+        "* See the support page: [Locate the project ID](https://support.google.com/googleapi/answer/7014113)"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "set_project_id"
+      },
+      "outputs": [],
+      "source": [
+        "PROJECT_ID = \"[your-project-id]\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# Set the project id\n",
+        "! gcloud config set project {PROJECT_ID}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "region"
+      },
+      "source": [
+        "#### Region\n",
+        "\n",
+        "You can also change the `REGION` variable used by Vertex AI. Learn more about [Vertex AI regions](https://cloud.google.com/vertex-ai/docs/general/locations)."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "region"
+      },
+      "outputs": [],
+      "source": [
+        "REGION = \"us-central1\"  # @param {type: \"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "gcp_authenticate"
+      },
+      "source": [
+        "### Authenticate your Google Cloud account\n",
+        "\n",
+        "Depending on your Jupyter environment, you may have to manually authenticate. Follow the relevant instructions below."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "FvQeFm3Gv5mR"
+      },
+      "source": [
+        "**1. Vertex AI Workbench**\n",
+        "* Do nothing as you are already authenticated."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "ad1138a125ea"
+      },
+      "source": [
+        "**2. Local JupyterLab instance, uncomment and run:**"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "ce6043da7b33"
+      },
+      "outputs": [],
+      "source": [
+        "# ! gcloud auth login"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0367eac06a10"
+      },
+      "source": [
+        "**3. Colab, uncomment and run:**"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "21ad4dbb4a61"
+      },
+      "outputs": [],
+      "source": [
+        "# from google.colab import auth\n",
+        "# auth.authenticate_user()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "c13224697bfb"
+      },
+      "source": [
+        "**4. Service account or other**\n",
+        "* See how to grant Cloud Storage permissions to your service account at https://cloud.google.com/storage/docs/gsutil/commands/iam#ch-examples."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "bucket:mbsdk"
+      },
+      "source": [
+        "### Create a Cloud Storage bucket\n",
+        "\n",
+        "Create a storage bucket to store intermediate artifacts such as datasets."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "bucket"
+      },
+      "outputs": [],
+      "source": [
+        "BUCKET_URI = f\"gs://your-bucket-name-unique-{PROJECT_ID}\"  # @param {type:\"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "create_bucket"
+      },
+      "source": [
+        "**Only if your bucket doesn't already exist**: Run the following cell to create your Cloud Storage bucket."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "09kHSsKmv5mT"
+      },
+      "outputs": [],
+      "source": [
+        "! gsutil mb -l $REGION -p $PROJECT_ID $BUCKET_URI"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "setup_vars"
+      },
+      "source": [
+        "### Set up variables\n",
+        "\n",
+        "Next, set up some variables used throughout the tutorial.\n",
+        "### Import libraries and define constants"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "import_aip:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "import google.cloud.aiplatform as aiplatform"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "init_aip:mbsdk"
+      },
+      "source": [
+        "## Initialize Vertex AI SDK for Python\n",
+        "\n",
+        "Initialize the Vertex AI SDK for Python for your project and corresponding bucket."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "init_aip:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "aiplatform.init(project=PROJECT_ID, staging_bucket=BUCKET_URI)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "tutorial_start:automl"
+      },
+      "source": [
+        "# Tutorial\n",
+        "\n",
+        "Now you are ready to start creating your own AutoML image classification model."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "import_file:u_dataset,csv"
+      },
+      "source": [
+        "#### Location of Cloud Storage training data.\n",
+        "\n",
+        "Now set the variable `IMPORT_FILE` to the location of the CSV index file in Cloud Storage."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "import_file:flowers,csv,icn"
+      },
+      "outputs": [],
+      "source": [
+        "IMPORT_FILE = (\n",
+        "    \"gs://cloud-samples-data/vision/automl_classification/flowers/all_data_v2.csv\"\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "quick_peek:csv"
+      },
+      "source": [
+        "#### Quick peek at your data\n",
+        "\n",
+        "This tutorial uses a version of the Flowers dataset that is stored in a public Cloud Storage bucket, using a CSV index file.\n",
+        "\n",
+        "Start by doing a quick peek at the data. You count the number of examples by counting the number of rows in the CSV index file  (`wc -l`) and then peek at the first few rows."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "quick_peek:csv"
+      },
+      "outputs": [],
+      "source": [
+        "if \"IMPORT_FILES\" in globals():\n",
+        "    FILE = IMPORT_FILES[0]\n",
+        "else:\n",
+        "    FILE = IMPORT_FILE\n",
+        "\n",
+        "count = ! gsutil cat $FILE | wc -l\n",
+        "print(\"Number of Examples\", int(count[0]))\n",
+        "\n",
+        "print(\"First 10 rows\")\n",
+        "! gsutil cat $FILE | head"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "create_dataset:image,icn"
+      },
+      "source": [
+        "### Create the Dataset\n",
+        "\n",
+        "Next, create the `Dataset` resource using the `create` method for the `ImageDataset` class, which takes the following parameters:\n",
+        "\n",
+        "- `display_name`: The human readable name for the `Dataset` resource.\n",
+        "- `gcs_source`: A list of one or more dataset index files to import the data items into the `Dataset` resource.\n",
+        "- `import_schema_uri`: The data labeling schema for the data items.\n",
+        "\n",
+        "This operation may take several minutes."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "create_dataset:image,icn"
+      },
+      "outputs": [],
+      "source": [
+        "dataset = aiplatform.ImageDataset.create(\n",
+        "    display_name=\"Flowers\",\n",
+        "    gcs_source=[IMPORT_FILE],\n",
+        "    import_schema_uri=aiplatform.schema.dataset.ioformat.image.single_label_classification,\n",
+        ")\n",
+        "\n",
+        "print(dataset.resource_name)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "create_automl_pipeline:image,icn"
+      },
+      "source": [
+        "### Create and run training pipeline\n",
+        "\n",
+        "To train an AutoML model, you perform two steps: 1) create a training pipeline, and 2) run the pipeline.\n",
+        "\n",
+        "#### Create training pipeline\n",
+        "\n",
+        "An AutoML training pipeline is created with the `AutoMLImageTrainingJob` class, with the following parameters:\n",
+        "\n",
+        "- `display_name`: The human readable name for the `TrainingJob` resource.\n",
+        "- `prediction_type`: The type task to train the model for.\n",
+        "  - `classification`: An image classification model.\n",
+        "  - `object_detection`: An image object detection model.\n",
+        "- `multi_label`: If a classification task, whether single (`False`) or multi-labeled (`True`).\n",
+        "- `model_type`: The type of model for deployment.\n",
+        "  - `CLOUD`: Deployment on Google Cloud\n",
+        "  - `CLOUD_HIGH_ACCURACY_1`: Optimized for accuracy over latency for deployment on Google Cloud.\n",
+        "  - `CLOUD_LOW_LATENCY_`: Optimized for latency over accuracy for deployment on Google Cloud.\n",
+        "  - `MOBILE_TF_VERSATILE_1`: Deployment on an edge device.\n",
+        "  - `MOBILE_TF_HIGH_ACCURACY_1`:Optimized for accuracy over latency for deployment on an edge device.\n",
+        "  - `MOBILE_TF_LOW_LATENCY_1`: Optimized for latency over accuracy for deployment on an edge device.\n",
+        "- `base_model`: (optional) Transfer learning from existing `Model` resource -- supported for image classification only.\n",
+        "\n",
+        "The instantiated object is the DAG (directed acyclic graph) for the training job."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "create_automl_pipeline:image,icn"
+      },
+      "outputs": [],
+      "source": [
+        "dag = aiplatform.AutoMLImageTrainingJob(\n",
+        "    display_name=\"flowers\",\n",
+        "    prediction_type=\"classification\",\n",
+        "    multi_label=False,\n",
+        "    model_type=\"CLOUD\",\n",
+        "    base_model=None,\n",
+        ")\n",
+        "\n",
+        "print(dag)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "run_automl_pipeline:image"
+      },
+      "source": [
+        "#### Run the training pipeline\n",
+        "\n",
+        "Next, you run the DAG to start the training job by invoking the method `run`, with the following parameters:\n",
+        "\n",
+        "- `dataset`: The `Dataset` resource to train the model.\n",
+        "- `model_display_name`: The human readable name for the trained model.\n",
+        "- `training_fraction_split`: The percentage of the dataset to use for training.\n",
+        "- `test_fraction_split`: The percentage of the dataset to use for test (holdout data).\n",
+        "- `validation_fraction_split`: The percentage of the dataset to use for validation.\n",
+        "- `budget_milli_node_hours`: (optional) Maximum training time specified in unit of millihours (1000 = hour).\n",
+        "- `disable_early_stopping`: If `True`, training maybe completed before using the entire budget if the service believes it cannot further improve on the model objective measurements.\n",
+        "\n",
+        "The `run` method when completed returns the `Model` resource.\n",
+        "\n",
+        "The execution of the training pipeline will take upto 20 minutes."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "run_automl_pipeline:image"
+      },
+      "outputs": [],
+      "source": [
+        "model = dag.run(\n",
+        "    dataset=dataset,\n",
+        "    model_display_name=\"flowers\",\n",
+        "    training_fraction_split=0.8,\n",
+        "    validation_fraction_split=0.1,\n",
+        "    test_fraction_split=0.1,\n",
+        "    budget_milli_node_hours=8000,\n",
+        "    disable_early_stopping=False,\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "evaluate_the_model:mbsdk"
+      },
+      "source": [
+        "## Review model evaluation scores\n",
+        "\n",
+        "After your model training has finished, you can review the evaluation scores for it using the `list_model_evaluations()` method. This method will return an iterator for each evaluation slice."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "evaluate_the_model:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "model_evaluations = model.list_model_evaluations()\n",
+        "\n",
+        "for model_evaluation in model_evaluations:\n",
+        "    print(model_evaluation.to_dict())"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "deploy_model:mbsdk,automatic"
+      },
+      "source": [
+        "## Deploy the model\n",
+        "\n",
+        "Next, deploy your model for online prediction. To deploy the model, you invoke the `deploy` method."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "deploy_model:mbsdk,automatic"
+      },
+      "outputs": [],
+      "source": [
+        "endpoint = model.deploy()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "make_prediction"
+      },
+      "source": [
+        "## Send an online prediction request\n",
+        "\n",
+        "Send an online prediction to your deployed model."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "get_test_item"
+      },
+      "source": [
+        "### Get test item\n",
+        "\n",
+        "You will use an arbitrary example out of the dataset as a test item. Don't be concerned that the example was likely used in training the model -- we just want to demonstrate how to make a prediction."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "get_test_item:automl,icn,csv"
+      },
+      "outputs": [],
+      "source": [
+        "test_item = !gsutil cat $IMPORT_FILE | head -n1\n",
+        "if len(str(test_item[0]).split(\",\")) == 3:\n",
+        "    _, test_item, test_label = str(test_item[0]).split(\",\")\n",
+        "else:\n",
+        "    test_item, test_label = str(test_item[0]).split(\",\")\n",
+        "\n",
+        "print(test_item, test_label)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "predict_request:mbsdk,icn"
+      },
+      "source": [
+        "### Make the prediction\n",
+        "\n",
+        "Now that your `Model` resource is deployed to an `Endpoint` resource, you can do online predictions by sending prediction requests to the Endpoint resource.\n",
+        "\n",
+        "#### Request\n",
+        "\n",
+        "Since in this example your test item is in a Cloud Storage bucket, you open and read the contents of the image using `tf.io.gfile.Gfile()`. To pass the test data to the prediction service, you encode the bytes into base64 -- which makes the content safe from modification while transmitting binary data over the network.\n",
+        "\n",
+        "The format of each instance is:\n",
+        "\n",
+        "    { 'content': { 'b64': base64_encoded_bytes } }\n",
+        "\n",
+        "Since the `predict()` method can take multiple items (instances), send your single test item as a list of one test item.\n",
+        "\n",
+        "#### Response\n",
+        "\n",
+        "The response from the `predict()` call is a Python dictionary with the following entries:\n",
+        "\n",
+        "- `ids`: The internal assigned unique identifiers for each prediction request.\n",
+        "- `displayNames`: The class names for each class label.\n",
+        "- `confidences`: The predicted confidence, between 0 and 1, per class label.\n",
+        "- `deployed_model_id`: The Vertex AI identifier for the deployed Model resource which did the predictions."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "predict_request:mbsdk,icn"
+      },
+      "outputs": [],
+      "source": [
+        "import base64\n",
+        "\n",
+        "import tensorflow as tf\n",
+        "\n",
+        "with tf.io.gfile.GFile(test_item, \"rb\") as f:\n",
+        "    content = f.read()\n",
+        "\n",
+        "# The format of each instance should conform to the deployed model's prediction input schema.\n",
+        "instances = [{\"content\": base64.b64encode(content).decode(\"utf-8\")}]\n",
+        "\n",
+        "prediction = endpoint.predict(instances=instances)\n",
+        "\n",
+        "print(prediction)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "undeploy_model:mbsdk"
+      },
+      "source": [
+        "## Undeploy the model\n",
+        "\n",
+        "When you are done doing predictions, you undeploy the model from the `Endpoint` resouce. This deprovisions all compute resources and ends billing for the deployed model."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "undeploy_model:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "endpoint.undeploy_all()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "cleanup:mbsdk"
+      },
+      "source": [
+        "# Cleaning up\n",
+        "\n",
+        "To clean up all Google Cloud resources used in this project, you can [delete the Google Cloud\n",
+        "project](https://cloud.google.com/resource-manager/docs/creating-managing-projects#shutting_down_projects) you used for the tutorial.\n",
+        "\n",
+        "Otherwise, you can delete the individual resources you created in this tutorial."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "cleanup:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "delete_bucket = False\n",
+        "\n",
+        "# Delete the dataset using the Vertex dataset object\n",
+        "dataset.delete()\n",
+        "\n",
+        "try:\n",
+        "    endpoint.undeploy_all()\n",
+        "    endpoint.delete()\n",
+        "except Exception as e:\n",
+        "    print(e)\n",
+        "\n",
+        "# Delete the model using the Vertex model object\n",
+        "model.delete()\n",
+        "\n",
+        "# Delete the AutoML trainig job\n",
+        "dag.delete()\n",
+        "\n",
+        "if delete_bucket or os.getenv(\"IS_TESTING\"):\n",
+        "    ! gsutil rm -r $BUCKET_URI"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "name": "automl_image_classification_online_online_prediction.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/official/automl/automl_image_object_detection_export_edge.ipynb

@@ -0,0 +1,835 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "copyright"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2021 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "title"
+      },
+      "source": [
+        "# AutoML training image object detection model for export to edge\n",
+        "\n",
+        "<table align=\"left\">\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_image_object_detection_export_edge.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_image_object_detection_export_edge.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/workbench/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/main/notebooks/official/automl//automl_image_object_detection_export_edge.ipynb\">\n",
+        "       <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "      Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>\n",
+        "<br/><br/><br/>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "overview:automl,export_edge"
+      },
+      "source": [
+        "## Overview\n",
+        "\n",
+        "\n",
+        "This tutorial demonstrates how to use the Vertex AI SDK to create image object detection models to export as an Edge model using an AutoML model."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "objective:automl,training,export_edge"
+      },
+      "source": [
+        "### Objective\n",
+        "\n",
+        "In this tutorial, you create an AutoML image object detection model from a Python script using the Vertex SDK, and then export the model as an Edge model in TFLite format. You can alternatively create models with AutoML using the `gcloud` command-line tool or online using the Cloud Console.\n",
+        "\n",
+        "This tutorial uses the following Google Cloud ML services:\n",
+        "\n",
+        "- Vertex AI `Datasets`\n",
+        "- AutoML Image\n",
+        "\n",
+        "The steps performed include:\n",
+        "\n",
+        "- Create a Vertex `Dataset` resource.\n",
+        "- Train the model.\n",
+        "- Export the `Edge` model from the `Model` resource to Cloud Storage.\n",
+        "- Download the model locally.\n",
+        "- Make a local prediction."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "dataset:salads,iod"
+      },
+      "source": [
+        "### Dataset\n",
+        "\n",
+        "The dataset used for this tutorial is the Salads category of the [OpenImages dataset](https://www.tensorflow.org/datasets/catalog/open_images_v4) from [TensorFlow Datasets](https://www.tensorflow.org/datasets/catalog/overview). This dataset does not require any feature engineering. The version of the dataset you will use in this tutorial is stored in a public Cloud Storage bucket. The trained model predicts the bounding box locations and corresponding type of salad items in an image from a class of five items: salad, seafood, tomato, baked goods, or cheese."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "costs"
+      },
+      "source": [
+        "### Costs\n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI\n",
+        "* Cloud Storage\n",
+        "\n",
+        "Learn about [Vertex AI\n",
+        "pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage\n",
+        "pricing](https://cloud.google.com/storage/pricing), and use the [Pricing\n",
+        "Calculator](https://cloud.google.com/products/calculator/)\n",
+        "to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "install_aip:mbsdk"
+      },
+      "source": [
+        "## Installation\n",
+        "\n",
+        "Install the latest version of Vertex AI SDK for Python."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "install_aip:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "import os\n",
+        "\n",
+        "# Google Cloud Notebook\n",
+        "if os.path.exists(\"/opt/deeplearning/metadata/env_version\"):\n",
+        "    USER_FLAG = \"--user\"\n",
+        "else:\n",
+        "    USER_FLAG = \"\"\n",
+        "\n",
+        "! pip3 install --upgrade google-cloud-aiplatform $USER_FLAG\n",
+        "\n",
+        "if os.environ[\"IS_TESTING\"]:\n",
+        "    ! pip3 install --upgrade tensorflow $USER_FLAG"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "restart"
+      },
+      "source": [
+        "### Colab only: Uncomment the following cell to restart the kernel"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "D-ZBOjErv5mM"
+      },
+      "outputs": [],
+      "source": [
+        "# Automatically restart kernel after installs so that your environment can access the new packages\n",
+        "# import IPython\n",
+        "\n",
+        "# app = IPython.Application.instance()\n",
+        "# app.kernel.do_shutdown(True)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "before_you_begin:nogpu"
+      },
+      "source": [
+        "## Before you begin"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "before_you_begin:nogpu"
+      },
+      "source": [
+        "### Set your project ID\n",
+        "\n",
+        "**If you don't know your project ID**, try the following:\n",
+        "* Run `gcloud config list`.\n",
+        "* Run `gcloud projects list`.\n",
+        "* See the support page: [Locate the project ID](https://support.google.com/googleapi/answer/7014113)"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "set_project_id"
+      },
+      "outputs": [],
+      "source": [
+        "PROJECT_ID = \"[your-project-id]\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# Set the project id\n",
+        "! gcloud config set project {PROJECT_ID}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "region"
+      },
+      "source": [
+        "#### Region\n",
+        "\n",
+        "You can also change the `REGION` variable used by Vertex AI. Learn more about [Vertex AI regions](https://cloud.google.com/vertex-ai/docs/general/locations)."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "region"
+      },
+      "outputs": [],
+      "source": [
+        "REGION = \"us-central1\"  # @param {type: \"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "gcp_authenticate"
+      },
+      "source": [
+        "### Authenticate your Google Cloud account\n",
+        "\n",
+        "Depending on your Jupyter environment, you may have to manually authenticate. Follow the relevant instructions below."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "FvQeFm3Gv5mR"
+      },
+      "source": [
+        "**1. Vertex AI Workbench**\n",
+        "* Do nothing as you are already authenticated."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "ad1138a125ea"
+      },
+      "source": [
+        "**2. Local JupyterLab instance, uncomment and run:**"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "ce6043da7b33"
+      },
+      "outputs": [],
+      "source": [
+        "# ! gcloud auth login"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0367eac06a10"
+      },
+      "source": [
+        "**3. Colab, uncomment and run:**"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "21ad4dbb4a61"
+      },
+      "outputs": [],
+      "source": [
+        "# from google.colab import auth\n",
+        "# auth.authenticate_user()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "c13224697bfb"
+      },
+      "source": [
+        "**4. Service account or other**\n",
+        "* See how to grant Cloud Storage permissions to your service account at https://cloud.google.com/storage/docs/gsutil/commands/iam#ch-examples."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "bucket:mbsdk"
+      },
+      "source": [
+        "### Create a Cloud Storage bucket\n",
+        "\n",
+        "Create a storage bucket to store intermediate artifacts such as datasets."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "bucket"
+      },
+      "outputs": [],
+      "source": [
+        "BUCKET_URI = f\"gs://your-bucket-name-unique-{PROJECT_ID}\"  # @param {type:\"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "create_bucket"
+      },
+      "source": [
+        "**Only if your bucket doesn't already exist**: Run the following cell to create your Cloud Storage bucket."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "create_bucket"
+      },
+      "outputs": [],
+      "source": [
+        "! gsutil mb -l $REGION $BUCKET_URI"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "setup_vars"
+      },
+      "source": [
+        "### Set up variables\n",
+        "\n",
+        "Next, set up some variables used throughout the tutorial.\n",
+        "### Import libraries and define constants"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "import_aip:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "import google.cloud.aiplatform as aiplatform"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "init_aip:mbsdk"
+      },
+      "source": [
+        "## Initialize Vertex AI SDK for Python\n",
+        "\n",
+        "Initialize the Vertex AI SDK for Python for your project and corresponding bucket."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "init_aip:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "aiplatform.init(project=PROJECT_ID, staging_bucket=BUCKET_URI)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "tutorial_start:automl"
+      },
+      "source": [
+        "# Tutorial\n",
+        "\n",
+        "Now you are ready to start creating your own AutoML image object detection model."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "import_file:u_dataset,csv"
+      },
+      "source": [
+        "#### Location of Cloud Storage training data.\n",
+        "\n",
+        "Now set the variable `IMPORT_FILE` to the location of the CSV index file in Cloud Storage."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "import_file:salads,csv,iod"
+      },
+      "outputs": [],
+      "source": [
+        "IMPORT_FILE = \"gs://cloud-samples-data/vision/salads.csv\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "quick_peek:csv"
+      },
+      "source": [
+        "#### Quick peek at your data\n",
+        "\n",
+        "This tutorial uses a version of the Salads dataset that is stored in a public Cloud Storage bucket, using a CSV index file.\n",
+        "\n",
+        "Start by doing a quick peek at the data. You count the number of examples by counting the number of rows in the CSV index file  (`wc -l`) and then peek at the first few rows."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "quick_peek:csv"
+      },
+      "outputs": [],
+      "source": [
+        "if \"IMPORT_FILES\" in globals():\n",
+        "    FILE = IMPORT_FILES[0]\n",
+        "else:\n",
+        "    FILE = IMPORT_FILE\n",
+        "\n",
+        "count = ! gsutil cat $FILE | wc -l\n",
+        "print(\"Number of Examples\", int(count[0]))\n",
+        "\n",
+        "print(\"First 10 rows\")\n",
+        "! gsutil cat $FILE | head"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "create_dataset:image,iod"
+      },
+      "source": [
+        "### Create the Dataset\n",
+        "\n",
+        "Next, create the `Dataset` resource using the `create` method for the `ImageDataset` class, which takes the following parameters:\n",
+        "\n",
+        "- `display_name`: The human readable name for the `Dataset` resource.\n",
+        "- `gcs_source`: A list of one or more dataset index files to import the data items into the `Dataset` resource.\n",
+        "- `import_schema_uri`: The data labeling schema for the data items.\n",
+        "\n",
+        "This operation may take several minutes."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "create_dataset:image,iod"
+      },
+      "outputs": [],
+      "source": [
+        "dataset = aiplatform.ImageDataset.create(\n",
+        "    display_name=\"Salads\",\n",
+        "    gcs_source=[IMPORT_FILE],\n",
+        "    import_schema_uri=aiplatform.schema.dataset.ioformat.image.bounding_box,\n",
+        ")\n",
+        "\n",
+        "print(dataset.resource_name)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "create_automl_pipeline:image,edge,iod"
+      },
+      "source": [
+        "### Create and run training pipeline\n",
+        "\n",
+        "To train an AutoML model, you perform two steps: 1) create a training pipeline, and 2) run the pipeline.\n",
+        "\n",
+        "#### Create training pipeline\n",
+        "\n",
+        "An AutoML training pipeline is created with the `AutoMLImageTrainingJob` class, with the following parameters:\n",
+        "\n",
+        "- `display_name`: The human readable name for the `TrainingJob` resource.\n",
+        "- `prediction_type`: The type task to train the model for.\n",
+        "  - `classification`: An image classification model.\n",
+        "  - `object_detection`: An image object detection model.\n",
+        "- `multi_label`: If a classification task, whether single (`False`) or multi-labeled (`True`).\n",
+        "- `model_type`: The type of model for deployment.\n",
+        "  - `CLOUD`: Deployment on Google Cloud\n",
+        "  - `CLOUD_HIGH_ACCURACY_1`: Optimized for accuracy over latency for deployment on Google Cloud.\n",
+        "  - `CLOUD_LOW_LATENCY_`: Optimized for latency over accuracy for deployment on Google Cloud.\n",
+        "  - `MOBILE_TF_VERSATILE_1`: Deployment on an edge device.\n",
+        "  - `MOBILE_TF_HIGH_ACCURACY_1`:Optimized for accuracy over latency for deployment on an edge device.\n",
+        "  - `MOBILE_TF_LOW_LATENCY_1`: Optimized for latency over accuracy for deployment on an edge device.\n",
+        "- `base_model`: (optional) Transfer learning from existing `Model` resource -- supported for image classification only.\n",
+        "\n",
+        "The instantiated object is the DAG (directed acyclic graph) for the training job."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "create_automl_pipeline:image,edge,iod"
+      },
+      "outputs": [],
+      "source": [
+        "dag = aiplatform.AutoMLImageTrainingJob(\n",
+        "    display_name=\"salads\",\n",
+        "    prediction_type=\"object_detection\",\n",
+        "    multi_label=False,\n",
+        "    model_type=\"MOBILE_TF_LOW_LATENCY_1\",\n",
+        "    base_model=None,\n",
+        ")\n",
+        "\n",
+        "print(dag)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "run_automl_pipeline:image"
+      },
+      "source": [
+        "#### Run the training pipeline\n",
+        "\n",
+        "Next, you run the DAG to start the training job by invoking the method `run`, with the following parameters:\n",
+        "\n",
+        "- `dataset`: The `Dataset` resource to train the model.\n",
+        "- `model_display_name`: The human readable name for the trained model.\n",
+        "- `training_fraction_split`: The percentage of the dataset to use for training.\n",
+        "- `test_fraction_split`: The percentage of the dataset to use for test (holdout data).\n",
+        "- `validation_fraction_split`: The percentage of the dataset to use for validation.\n",
+        "- `budget_milli_node_hours`: (optional) Maximum training time specified in unit of millihours (1000 = hour).\n",
+        "- `disable_early_stopping`: If `True`, training maybe completed before using the entire budget if the service believes it cannot further improve on the model objective measurements.\n",
+        "\n",
+        "The `run` method when completed returns the `Model` resource.\n",
+        "\n",
+        "The execution of the training pipeline will take upto 60 minutes."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "run_automl_pipeline:image"
+      },
+      "outputs": [],
+      "source": [
+        "model = dag.run(\n",
+        "    dataset=dataset,\n",
+        "    model_display_name=\"salads\",\n",
+        "    training_fraction_split=0.8,\n",
+        "    validation_fraction_split=0.1,\n",
+        "    test_fraction_split=0.1,\n",
+        "    budget_milli_node_hours=20000,\n",
+        "    disable_early_stopping=False,\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "evaluate_the_model:mbsdk"
+      },
+      "source": [
+        "## Review model evaluation scores\n",
+        "\n",
+        "After your model training has finished, you can review the evaluation scores for it using the `list_model_evaluations()` method. This method will return an iterator for each evaluation slice."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "evaluate_the_model:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "model_evaluations = model.list_model_evaluations()\n",
+        "\n",
+        "for model_evaluation in model_evaluations:\n",
+        "    print(model_evaluation.to_dict())"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "export_model:mbsdk,image"
+      },
+      "source": [
+        "## Export as Edge model\n",
+        "\n",
+        "You can export an AutoML image object detection model as a `Edge` model which you can then custom deploy to an edge device or download locally. Use the method `export_model()` to export the model to Cloud Storage, which takes the following parameters:\n",
+        "\n",
+        "- `artifact_destination`: The Cloud Storage location to store the SavedFormat model artifacts to.\n",
+        "- `export_format_id`: The format to save the model format as. For AutoML image object detection there is just one option:\n",
+        "   - `tf-saved-model`: TensorFlow SavedFormat for deployment to a container.\n",
+        "   - `tflite`: TensorFlow Lite for deployment to an edge or mobile device.\n",
+        "   - `edgetpu-tflite`: TensorFlow Lite for TPU\n",
+        "   - `tf-js`: TensorFlow for web client\n",
+        "   - `coral-ml`: for Coral devices\n",
+        "\n",
+        "- `sync`: Whether to perform operational sychronously or asynchronously."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "export_model:mbsdk,image"
+      },
+      "outputs": [],
+      "source": [
+        "response = model.export_model(\n",
+        "    artifact_destination=BUCKET_URI, export_format_id=\"tflite\", sync=True\n",
+        ")\n",
+        "\n",
+        "model_package = response[\"artifactOutputUri\"]"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "download_model_artifacts:tflite"
+      },
+      "source": [
+        "#### Download the TFLite model artifacts\n",
+        "\n",
+        "Now that you have an exported TFLite version of your model, you can test the exported model locally, but first downloading it from Cloud Storage."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "download_model_artifacts:tflite"
+      },
+      "outputs": [],
+      "source": [
+        "! gsutil ls $model_package\n",
+        "# Download the model artifacts\n",
+        "! gsutil cp -r $model_package tflite\n",
+        "\n",
+        "tflite_path = \"tflite/model.tflite\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "instantiate_tflite_interpreter"
+      },
+      "source": [
+        "#### Instantiate a TFLite interpreter\n",
+        "\n",
+        "The TFLite version of the model is not a TensorFlow SavedModel format. You cannot directly use methods like predict(). Instead, one uses the TFLite interpreter. You must first setup the interpreter for the TFLite model as follows:\n",
+        "\n",
+        "- Instantiate an TFLite interpreter for the TFLite model.\n",
+        "- Instruct the interpreter to allocate input and output tensors for the model.\n",
+        "- Get detail information about the models input and output tensors that will need to be known for prediction."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "instantiate_tflite_interpreter"
+      },
+      "outputs": [],
+      "source": [
+        "import tensorflow as tf\n",
+        "\n",
+        "interpreter = tf.lite.Interpreter(model_path=tflite_path)\n",
+        "interpreter.allocate_tensors()\n",
+        "\n",
+        "input_details = interpreter.get_input_details()\n",
+        "output_details = interpreter.get_output_details()\n",
+        "input_shape = input_details[0][\"shape\"]\n",
+        "\n",
+        "print(\"input tensor shape\", input_shape)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "get_test_item"
+      },
+      "source": [
+        "### Get test item\n",
+        "\n",
+        "You will use an arbitrary example out of the dataset as a test item. Don't be concerned that the example was likely used in training the model -- we just want to demonstrate how to make a prediction."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "get_test_item:image,224x224"
+      },
+      "outputs": [],
+      "source": [
+        "test_items = ! gsutil cat $IMPORT_FILE | head -n1\n",
+        "test_item = test_items[0].split(\",\")[0]\n",
+        "\n",
+        "with tf.io.gfile.GFile(test_item, \"rb\") as f:\n",
+        "    content = f.read()\n",
+        "test_image = tf.io.decode_jpeg(content)\n",
+        "print(\"test image shape\", test_image.shape)\n",
+        "\n",
+        "test_image = tf.image.resize(test_image, (192, 192))\n",
+        "print(\"test image shape\", test_image.shape, test_image.dtype)\n",
+        "\n",
+        "test_image = tf.cast(test_image, dtype=tf.uint8).numpy()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "invoke_tflite_interpreter"
+      },
+      "source": [
+        "#### Make a prediction with TFLite model\n",
+        "\n",
+        "Finally, you do a prediction using your TFLite model, as follows:\n",
+        "\n",
+        "- Convert the test image into a batch of a single image (`np.expand_dims`)\n",
+        "- Set the input tensor for the interpreter to your batch of a single image (`data`).\n",
+        "- Invoke the interpreter.\n",
+        "- Retrieve the softmax probabilities for the prediction (`get_tensor`).\n",
+        "- Determine which label had the highest probability (`np.argmax`)."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "invoke_tflite_interpreter"
+      },
+      "outputs": [],
+      "source": [
+        "import numpy as np\n",
+        "\n",
+        "data = np.expand_dims(test_image, axis=0)\n",
+        "\n",
+        "interpreter.set_tensor(input_details[0][\"index\"], data)\n",
+        "\n",
+        "interpreter.invoke()\n",
+        "\n",
+        "softmax = interpreter.get_tensor(output_details[0][\"index\"])\n",
+        "\n",
+        "label = np.argmax(softmax)\n",
+        "\n",
+        "print(label)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "cleanup:mbsdk"
+      },
+      "source": [
+        "# Cleaning up\n",
+        "\n",
+        "To clean up all Google Cloud resources used in this project, you can [delete the Google Cloud\n",
+        "project](https://cloud.google.com/resource-manager/docs/creating-managing-projects#shutting_down_projects) you used for the tutorial.\n",
+        "\n",
+        "Otherwise, you can delete the individual resources you created in this tutorial."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "cleanup:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "delete_bucket = False\n",
+        "\n",
+        "# Delete the dataset using the Vertex dataset object\n",
+        "dataset.delete()\n",
+        "\n",
+        "# Delete the model using the Vertex model object\n",
+        "model.delete()\n",
+        "\n",
+        "# Delete the AutoML trainig job\n",
+        "dag.delete()\n",
+        "\n",
+        "if delete_bucket or os.getenv(\"IS_TESTING\"):\n",
+        "    ! gsutil rm -r $BUCKET_URI"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "name": "automl_image_object_detection_export_edge.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/official/automl/automl_image_object_detection_online_prediction.ipynb

@@ -0,0 +1,815 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "copyright"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2021 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "title"
+      },
+      "source": [
+        "# AutoML training image object detection model for online prediction\n",
+        "\n",
+        "<table align=\"left\">\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_image_object_detection_online_prediction.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_image_object_detection_online_prediction.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/workbench/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/main/notebooks/official/automl/automl_image_object_detection_online_prediction.ipynb\">\n",
+        "       <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "      Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>\n",
+        "<br/><br/><br/>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "overview:automl"
+      },
+      "source": [
+        "## Overview\n",
+        "\n",
+        "\n",
+        "This tutorial demonstrates how to use the Vertex AI SDK to create image object detection models and do online prediction using a Google Cloud [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) model.\n",
+        "\n",
+        "Learn more about [Object detection for image data](https://cloud.google.com/vertex-ai/docs/training-overview#object_detection_for_images)."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "objective:automl,training,online_prediction"
+      },
+      "source": [
+        "### Objective\n",
+        "\n",
+        "In this tutorial, you create an AutoML image object detection model and deploy for online prediction from a Python script using the Vertex AI SDK. You can alternatively create and deploy models using the `gcloud` command-line tool or online using the Cloud Console.\n",
+        "\n",
+        "This tutorial uses the following Google Cloud ML services:\n",
+        "\n",
+        "- Vertex AI `Datasets`\n",
+        "- AutoML Image\n",
+        "- Vertex AI `Model Registry`\n",
+        "- Vertex AI `Predictions`\n",
+        "\n",
+        "The steps performed include:\n",
+        "\n",
+        "- Create a Vertex `Dataset` resource.\n",
+        "- Train the model.\n",
+        "- View the model evaluation.\n",
+        "- Deploy the `Model` resource to a serving `Endpoint` resource.\n",
+        "- Make a prediction.\n",
+        "- Undeploy the `Model`."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "dataset:salads,iod"
+      },
+      "source": [
+        "### Dataset\n",
+        "\n",
+        "The dataset used for this tutorial is the Salads category of the [OpenImages dataset](https://www.tensorflow.org/datasets/catalog/open_images_v4) from [TensorFlow Datasets](https://www.tensorflow.org/datasets/catalog/overview). This dataset does not require any feature engineering. The version of the dataset you will use in this tutorial is stored in a public Cloud Storage bucket. The trained model predicts the bounding box locations and corresponding type of salad items in an image from a class of five items: salad, seafood, tomato, baked goods, or cheese."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "costs"
+      },
+      "source": [
+        "### Costs\n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI\n",
+        "* Cloud Storage\n",
+        "\n",
+        "Learn about [Vertex AI\n",
+        "pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage\n",
+        "pricing](https://cloud.google.com/storage/pricing), and use the [Pricing\n",
+        "Calculator](https://cloud.google.com/products/calculator/)\n",
+        "to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "install_aip:mbsdk"
+      },
+      "source": [
+        "## Installation\n",
+        "\n",
+        "Install the latest version of Vertex AI SDK for Python."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "install_aip:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "import os\n",
+        "\n",
+        "USER_FLAG = \"\"\n",
+        "\n",
+        "! pip3 install --upgrade google-cloud-aiplatform \\\n",
+        "                         tensorflow $USER_FLAG\n",
+        "\n",
+        "if os.environ[\"IS_TESTING\"]:\n",
+        "    ! pip3 install --upgrade tensorflow $USER_FLAG"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "restart"
+      },
+      "source": [
+        "### Colab only: Uncomment the following cell to restart the kernel"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "D-ZBOjErv5mM"
+      },
+      "outputs": [],
+      "source": [
+        "# Automatically restart kernel after installs so that your environment can access the new packages\n",
+        "# import IPython\n",
+        "\n",
+        "# app = IPython.Application.instance()\n",
+        "# app.kernel.do_shutdown(True)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "before_you_begin:nogpu"
+      },
+      "source": [
+        "## Before you begin"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "before_you_begin:nogpu"
+      },
+      "source": [
+        "### Set your project ID\n",
+        "\n",
+        "**If you don't know your project ID**, try the following:\n",
+        "* Run `gcloud config list`.\n",
+        "* Run `gcloud projects list`.\n",
+        "* See the support page: [Locate the project ID](https://support.google.com/googleapi/answer/7014113)"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "set_project_id"
+      },
+      "outputs": [],
+      "source": [
+        "PROJECT_ID = \"[your-project-id]\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# Set the project id\n",
+        "! gcloud config set project {PROJECT_ID}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "region"
+      },
+      "source": [
+        "#### Region\n",
+        "\n",
+        "You can also change the `REGION` variable used by Vertex AI. Learn more about [Vertex AI regions](https://cloud.google.com/vertex-ai/docs/general/locations)."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "region"
+      },
+      "outputs": [],
+      "source": [
+        "REGION = \"us-central1\"  # @param {type: \"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "gcp_authenticate"
+      },
+      "source": [
+        "### Authenticate your Google Cloud account\n",
+        "\n",
+        "Depending on your Jupyter environment, you may have to manually authenticate. Follow the relevant instructions below."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "FvQeFm3Gv5mR"
+      },
+      "source": [
+        "**1. Vertex AI Workbench**\n",
+        "* Do nothing as you are already authenticated."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "ad1138a125ea"
+      },
+      "source": [
+        "**2. Local JupyterLab instance, uncomment and run:**"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "ce6043da7b33"
+      },
+      "outputs": [],
+      "source": [
+        "# ! gcloud auth login"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0367eac06a10"
+      },
+      "source": [
+        "**3. Colab, uncomment and run:**"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "21ad4dbb4a61"
+      },
+      "outputs": [],
+      "source": [
+        "# from google.colab import auth\n",
+        "# auth.authenticate_user()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "c13224697bfb"
+      },
+      "source": [
+        "**4. Service account or other**\n",
+        "* See how to grant Cloud Storage permissions to your service account at https://cloud.google.com/storage/docs/gsutil/commands/iam#ch-examples."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "bucket:mbsdk"
+      },
+      "source": [
+        "### Create a Cloud Storage bucket\n",
+        "\n",
+        "Create a storage bucket to store intermediate artifacts such as datasets."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "bucket"
+      },
+      "outputs": [],
+      "source": [
+        "BUCKET_URI = f\"gs://your-bucket-name-unique-{PROJECT_ID}\"  # @param {type:\"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "create_bucket"
+      },
+      "source": [
+        "**Only if your bucket doesn't already exist**: Run the following cell to create your Cloud Storage bucket."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "create_bucket"
+      },
+      "outputs": [],
+      "source": [
+        "! gsutil mb -l $REGION $BUCKET_URI"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "setup_vars"
+      },
+      "source": [
+        "### Set up variables\n",
+        "\n",
+        "Next, set up some variables used throughout the tutorial.\n",
+        "### Import libraries and define constants"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "import_aip:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "import google.cloud.aiplatform as aiplatform"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "init_aip:mbsdk"
+      },
+      "source": [
+        "## Initialize Vertex AI SDK for Python\n",
+        "\n",
+        "Initialize the Vertex AI SDK for Python for your project and corresponding bucket."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "init_aip:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "aiplatform.init(project=PROJECT_ID, staging_bucket=BUCKET_URI)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "tutorial_start:automl"
+      },
+      "source": [
+        "# Tutorial\n",
+        "\n",
+        "Now you are ready to start creating your own AutoML image object detection model."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "import_file:u_dataset,csv"
+      },
+      "source": [
+        "#### Location of Cloud Storage training data.\n",
+        "\n",
+        "Now set the variable `IMPORT_FILE` to the location of the CSV index file in Cloud Storage."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "import_file:salads,csv,iod"
+      },
+      "outputs": [],
+      "source": [
+        "IMPORT_FILE = \"gs://cloud-samples-data/vision/salads.csv\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "quick_peek:csv"
+      },
+      "source": [
+        "#### Quick peek at your data\n",
+        "\n",
+        "This tutorial uses a version of the Salads dataset that is stored in a public Cloud Storage bucket, using a CSV index file.\n",
+        "\n",
+        "Start by doing a quick peek at the data. You count the number of examples by counting the number of rows in the CSV index file  (`wc -l`) and then peek at the first few rows."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "quick_peek:csv"
+      },
+      "outputs": [],
+      "source": [
+        "if \"IMPORT_FILES\" in globals():\n",
+        "    FILE = IMPORT_FILES[0]\n",
+        "else:\n",
+        "    FILE = IMPORT_FILE\n",
+        "\n",
+        "count = ! gsutil cat $FILE | wc -l\n",
+        "print(\"Number of Examples\", int(count[0]))\n",
+        "\n",
+        "print(\"First 10 rows\")\n",
+        "! gsutil cat $FILE | head"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "create_dataset:image,iod"
+      },
+      "source": [
+        "### Create the Dataset\n",
+        "\n",
+        "Next, create the `Dataset` resource using the `create` method for the `ImageDataset` class, which takes the following parameters:\n",
+        "\n",
+        "- `display_name`: The human readable name for the `Dataset` resource.\n",
+        "- `gcs_source`: A list of one or more dataset index files to import the data items into the `Dataset` resource.\n",
+        "- `import_schema_uri`: The data labeling schema for the data items.\n",
+        "\n",
+        "This operation may take several minutes."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "create_dataset:image,iod"
+      },
+      "outputs": [],
+      "source": [
+        "dataset = aiplatform.ImageDataset.create(\n",
+        "    display_name=\"Salads\",\n",
+        "    gcs_source=[IMPORT_FILE],\n",
+        "    import_schema_uri=aiplatform.schema.dataset.ioformat.image.bounding_box,\n",
+        ")\n",
+        "\n",
+        "print(dataset.resource_name)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "create_automl_pipeline:image,iod"
+      },
+      "source": [
+        "### Create and run training pipeline\n",
+        "\n",
+        "To train an AutoML model, you perform two steps: 1) create a training pipeline, and 2) run the pipeline.\n",
+        "\n",
+        "#### Create training pipeline\n",
+        "\n",
+        "An AutoML training pipeline is created with the `AutoMLImageTrainingJob` class, with the following parameters:\n",
+        "\n",
+        "- `display_name`: The human readable name for the `TrainingJob` resource.\n",
+        "- `prediction_type`: The type task to train the model for.\n",
+        "  - `classification`: An image classification model.\n",
+        "  - `object_detection`: An image object detection model.\n",
+        "- `multi_label`: If a classification task, whether single (`False`) or multi-labeled (`True`).\n",
+        "- `model_type`: The type of model for deployment.\n",
+        "  - `CLOUD`: Deployment on Google Cloud\n",
+        "  - `CLOUD_HIGH_ACCURACY_1`: Optimized for accuracy over latency for deployment on Google Cloud.\n",
+        "  - `CLOUD_LOW_LATENCY_`: Optimized for latency over accuracy for deployment on Google Cloud.\n",
+        "  - `MOBILE_TF_VERSATILE_1`: Deployment on an edge device.\n",
+        "  - `MOBILE_TF_HIGH_ACCURACY_1`:Optimized for accuracy over latency for deployment on an edge device.\n",
+        "  - `MOBILE_TF_LOW_LATENCY_1`: Optimized for latency over accuracy for deployment on an edge device.\n",
+        "- `base_model`: (optional) Transfer learning from existing `Model` resource -- supported for image classification only.\n",
+        "\n",
+        "The instantiated object is the DAG (directed acyclic graph) for the training job."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "create_automl_pipeline:image,iod"
+      },
+      "outputs": [],
+      "source": [
+        "dag = aiplatform.AutoMLImageTrainingJob(\n",
+        "    display_name=\"salads\",\n",
+        "    prediction_type=\"object_detection\",\n",
+        "    multi_label=False,\n",
+        "    model_type=\"CLOUD\",\n",
+        "    base_model=None,\n",
+        ")\n",
+        "\n",
+        "print(dag)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "run_automl_pipeline:image"
+      },
+      "source": [
+        "#### Run the training pipeline\n",
+        "\n",
+        "Next, you run the DAG to start the training job by invoking the method `run`, with the following parameters:\n",
+        "\n",
+        "- `dataset`: The `Dataset` resource to train the model.\n",
+        "- `model_display_name`: The human readable name for the trained model.\n",
+        "- `training_fraction_split`: The percentage of the dataset to use for training.\n",
+        "- `test_fraction_split`: The percentage of the dataset to use for test (holdout data).\n",
+        "- `validation_fraction_split`: The percentage of the dataset to use for validation.\n",
+        "- `budget_milli_node_hours`: (optional) Maximum training time specified in unit of millihours (1000 = hour).\n",
+        "- `disable_early_stopping`: If `True`, training maybe completed before using the entire budget if the service believes it cannot further improve on the model objective measurements.\n",
+        "\n",
+        "The `run` method when completed returns the `Model` resource.\n",
+        "\n",
+        "The execution of the training pipeline will take upto 60 minutes."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "run_automl_pipeline:image"
+      },
+      "outputs": [],
+      "source": [
+        "model = dag.run(\n",
+        "    dataset=dataset,\n",
+        "    model_display_name=\"salads\",\n",
+        "    training_fraction_split=0.8,\n",
+        "    validation_fraction_split=0.1,\n",
+        "    test_fraction_split=0.1,\n",
+        "    budget_milli_node_hours=20000,\n",
+        "    disable_early_stopping=False,\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "evaluate_the_model:mbsdk"
+      },
+      "source": [
+        "## Review model evaluation scores\n",
+        "\n",
+        "After your model training has finished, you can review the evaluation scores for it using the `list_model_evaluations()` method. This method will return an iterator for each evaluation slice."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "evaluate_the_model:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "model_evaluations = model.list_model_evaluations()\n",
+        "\n",
+        "for model_evaluation in model_evaluations:\n",
+        "    print(model_evaluation.to_dict())"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "deploy_model:mbsdk,automatic"
+      },
+      "source": [
+        "## Deploy the model\n",
+        "\n",
+        "Next, deploy your model for online prediction. To deploy the model, you invoke the `deploy` method."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "deploy_model:mbsdk,automatic"
+      },
+      "outputs": [],
+      "source": [
+        "endpoint = model.deploy()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "make_prediction"
+      },
+      "source": [
+        "## Send an online prediction request\n",
+        "\n",
+        "Send an online prediction to your deployed model."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "get_test_item"
+      },
+      "source": [
+        "### Get test item\n",
+        "\n",
+        "You will use an arbitrary example out of the dataset as a test item. Don't be concerned that the example was likely used in training the model -- we just want to demonstrate how to make a prediction."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "get_test_item:automl,iod,csv"
+      },
+      "outputs": [],
+      "source": [
+        "test_items = !gsutil cat $IMPORT_FILE | head -n1\n",
+        "cols = str(test_items[0]).split(\",\")\n",
+        "if len(cols) == 11:\n",
+        "    test_item = str(cols[1])\n",
+        "    test_label = str(cols[2])\n",
+        "else:\n",
+        "    test_item = str(cols[0])\n",
+        "    test_label = str(cols[1])\n",
+        "\n",
+        "print(test_item, test_label)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "predict_request:mbsdk,iod"
+      },
+      "source": [
+        "### Make the prediction\n",
+        "\n",
+        "Now that your `Model` resource is deployed to an `Endpoint` resource, you can do online predictions by sending prediction requests to the Endpoint resource.\n",
+        "\n",
+        "#### Request\n",
+        "\n",
+        "Since in this example your test item is in a Cloud Storage bucket, you open and read the contents of the image using `tf.io.gfile.Gfile()`. To pass the test data to the prediction service, you encode the bytes into base64 -- which makes the content safe from modification while transmitting binary data over the network.\n",
+        "\n",
+        "The format of each instance is:\n",
+        "\n",
+        "    { 'content': { 'b64': base64_encoded_bytes } }\n",
+        "\n",
+        "Since the `predict()` method can take multiple items (instances), send your single test item as a list of one test item.\n",
+        "\n",
+        "#### Response\n",
+        "\n",
+        "The response from the `predict()` call is a Python dictionary with the following entries:\n",
+        "\n",
+        "- `ids`: The internal assigned unique identifiers for each prediction request.\n",
+        "- `displayNames`: The class names for each class label.\n",
+        "- `confidences`: The predicted confidence, between 0 and 1, per class label.\n",
+        "- `bboxes`: The bounding box of each detected object.\n",
+        "- `deployed_model_id`: The Vertex AI identifier for the deployed Model resource which did the predictions."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "predict_request:mbsdk,iod"
+      },
+      "outputs": [],
+      "source": [
+        "import base64\n",
+        "\n",
+        "import tensorflow as tf\n",
+        "\n",
+        "with tf.io.gfile.GFile(test_item, \"rb\") as f:\n",
+        "    content = f.read()\n",
+        "\n",
+        "# The format of each instance should conform to the deployed model's prediction input schema.\n",
+        "instances = [{\"content\": base64.b64encode(content).decode(\"utf-8\")}]\n",
+        "\n",
+        "prediction = endpoint.predict(instances=instances)\n",
+        "\n",
+        "print(prediction)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "undeploy_model:mbsdk"
+      },
+      "source": [
+        "## Undeploy the model\n",
+        "\n",
+        "When you are done doing predictions, you undeploy the model from the `Endpoint` resource. This deprovisions all compute resources and ends billing for the deployed model."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "undeploy_model:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "endpoint.undeploy_all()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "cleanup:mbsdk"
+      },
+      "source": [
+        "# Cleaning up\n",
+        "\n",
+        "To clean up all Google Cloud resources used in this project, you can [delete the Google Cloud\n",
+        "project](https://cloud.google.com/resource-manager/docs/creating-managing-projects#shutting_down_projects) you used for the tutorial.\n",
+        "\n",
+        "Otherwise, you can delete the individual resources you created in this tutorial."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "cleanup:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "import os\n",
+        "\n",
+        "delete_bucket = False\n",
+        "\n",
+        "# Delete the dataset using the Vertex dataset object\n",
+        "dataset.delete()\n",
+        "\n",
+        "try:\n",
+        "    endpoint.undeploy_all()\n",
+        "    endpoint.delete()\n",
+        "except Exception as e:\n",
+        "    print(e)\n",
+        "\n",
+        "# Delete the model using the Vertex model object\n",
+        "model.delete()\n",
+        "\n",
+        "# Delete the AutoML trainig job\n",
+        "dag.delete()\n",
+        "\n",
+        "if delete_bucket or os.getenv(\"IS_TESTING\"):\n",
+        "    ! gsutil rm -r $BUCKET_URI"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "name": "automl_image_object_detection_online_prediction.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/official/automl/automl_tabular_on_vertex_pipelines.ipynb

@@ -61,7 +61,9 @@
       "source": [
         "## Overview\n",
         "\n",
-        "In this tutorial, you will use two Vertex AI Tabular Workflows pipelines to train AutoML models using different configurations. You will see how `get_automl_tabular_pipeline_and_parameters` gives you the ability to customize the default AutoML Tabular pipeline, and how `get_skip_architecture_search_pipeline_and_parameters` allows you to reduce the training time and cost for an AutoML model by using the tuning results from a previous pipeline run."
+        "In this tutorial, you will use two Vertex AI Tabular Workflows pipelines to train AutoML models using different configurations. You will see how `get_automl_tabular_pipeline_and_parameters` gives you the ability to customize the default AutoML Tabular pipeline, and how `get_skip_architecture_search_pipeline_and_parameters` allows you to reduce the training time and cost for an AutoML model by using the tuning results from a previous pipeline run.\n",
+        "\n",
+        "Learn more about [Tabular Workflow for E2E AutoML](https://cloud.google.com/vertex-ai/docs/tabular-data/tabular-workflows/e2e-automl)."
       ]
     },
     {
@@ -168,7 +170,7 @@
       },
       "outputs": [],
       "source": [
-        "!pip install -U google-cloud-pipeline-components -q"
+        "!pip install -U google-cloud-pipeline-components==1.0.25 -q"
       ]
     },
     {
@@ -331,7 +333,10 @@
       },
       "outputs": [],
       "source": [
-        "REGION = \"[your-region]\"  # @param {type: \"string\"}"
+        "REGION = \"us-central1\"  # @param {type: \"string\"}\n",
+        "\n",
+        "if REGION == \"[your-region]\":\n",
+        "    REGION = \"us-central1\""
       ]
     },
     {
@@ -377,8 +382,11 @@
         "import os\n",
         "import sys\n",
         "\n",
-        "# If on Google Cloud Notebook, then don't execute this code\n",
-        "if not os.path.exists(\"/opt/deeplearning/metadata/env_version\"):\n",
+        "# If on Vertex AI Workbench, then don't execute this code\n",
+        "IS_COLAB = \"google.colab\" in sys.modules\n",
+        "if not os.path.exists(\"/opt/deeplearning/metadata/env_version\") and not os.getenv(\n",
+        "    \"DL_ANACONDA_HOME\"\n",
+        "):\n",
         "    if \"google.colab\" in sys.modules:\n",
         "        from google.colab import auth as google_auth\n",
         "\n",
@@ -386,10 +394,9 @@
         "\n",
         "    # If you are running this notebook locally, replace the string below with the\n",
         "    # path to your service account key and run this cell to authenticate your GCP\n",
-        "    # account. Alternatively, you may edit this notebook to authenticate using\n",
-        "    # gcloud.\n",
+        "    # account.\n",
         "    elif not os.getenv(\"IS_TESTING\"):\n",
-        "        %env GOOGLE_APPLICATION_CREDENTIALS ''"
+        "        %env GOOGLE_APPLICATION_CREDENTIALS '[your-service-account-key-path]'"
       ]
     },
     {
@@ -476,6 +483,77 @@
         "! gsutil ls -al $BUCKET_URI"
       ]
     },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "44accda192d5"
+      },
+      "source": [
+        "#### Service Account\n",
+        "\n",
+        "You use a service account to create Vertex AI Pipeline jobs. If you do not want to use your project's Compute Engine service account, set `SERVICE_ACCOUNT` to another service account ID."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "e0c9c4f84849"
+      },
+      "outputs": [],
+      "source": [
+        "SERVICE_ACCOUNT = \"[your-service-account]\""
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "604ae09ab6d3"
+      },
+      "outputs": [],
+      "source": [
+        "if (\n",
+        "    SERVICE_ACCOUNT == \"\"\n",
+        "    or SERVICE_ACCOUNT is None\n",
+        "    or SERVICE_ACCOUNT == \"[your-service-account]\"\n",
+        "):\n",
+        "    # Get your service account from gcloud\n",
+        "    if not IS_COLAB:\n",
+        "        shell_output = !gcloud auth list 2>/dev/null\n",
+        "        SERVICE_ACCOUNT = shell_output[2].replace(\"*\", \"\").strip()\n",
+        "\n",
+        "    else:  # IS_COLAB:\n",
+        "        shell_output = ! gcloud projects describe  $PROJECT_ID\n",
+        "        project_number = shell_output[-1].split(\":\")[1].strip().replace(\"'\", \"\")\n",
+        "        SERVICE_ACCOUNT = f\"{project_number}-compute@developer.gserviceaccount.com\"\n",
+        "\n",
+        "    print(\"Service Account:\", SERVICE_ACCOUNT)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "d1ecb60964d5"
+      },
+      "source": [
+        "#### Set service account access for Vertex AI Pipelines\n",
+        "Run the following commands to grant your service account access to read and write pipeline artifacts in the bucket that you created in the previous step. You only need to run this step once per service account."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "a592f0a380c2"
+      },
+      "outputs": [],
+      "source": [
+        "! gsutil iam ch serviceAccount:{SERVICE_ACCOUNT}:roles/storage.objectCreator $BUCKET_URI\n",
+        "\n",
+        "! gsutil iam ch serviceAccount:{SERVICE_ACCOUNT}:roles/storage.objectViewer $BUCKET_URI"
+      ]
+    },
     {
       "cell_type": "markdown",
       "metadata": {
@@ -834,6 +912,7 @@
         "\n",
         "job.run()\n",
         "\n",
+        "\n",
         "pipeline_task_details = job.gca_resource.job_detail.task_details\n",
         "\n",
         "if export_additional_model_without_custom_ops:\n",
@@ -875,6 +954,7 @@
         "stage_1_tuner_task = get_task_detail(\n",
         "    pipeline_task_details, \"automl-tabular-stage-1-tuner\"\n",
         ")\n",
+        "\n",
         "stage_1_tuning_result_artifact_uri = (\n",
         "    stage_1_tuner_task.outputs[\"tuning_result_output\"].artifacts[0].uri\n",
         ")"

notebooks/official/automl/automl_text_entity_extraction_batch_prediction.ipynb

@@ -0,0 +1,818 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "copyright"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2021 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "title"
+      },
+      "source": [
+        "# AutoML training text entity extraction model for batch prediction\n",
+        "\n",
+        "<table align=\"left\">\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_text_entity_extraction_batch_prediction.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_text_entity_extraction_batch_prediction.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/workbench/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/main/notebooks/official/automl/automl_text_entity_extraction_batch_prediction.ipynb\">\n",
+        "        <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "      Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>\n",
+        "<br/><br/><br/>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "overview:automl"
+      },
+      "source": [
+        "## Overview\n",
+        "\n",
+        "\n",
+        "This tutorial demonstrates how to use the Vertex AI SDK to create text entity extraction models and do batch prediction using a Google Cloud [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) model.\n",
+        "\n",
+        "Learn more about [Entity extraction for text data](https://cloud.google.com/vertex-ai/docs/training-overview#entity_extraction_for_text)."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "objective:automl,training,batch_prediction"
+      },
+      "source": [
+        "### Objective\n",
+        "\n",
+        "In this tutorial, you create an AutoML text entity extraction model from a Python script, and then do a batch prediction using the Vertex AI SDK. You can alternatively create and deploy models using the `gcloud` command-line tool or online using the Cloud Console.\n",
+        "\n",
+        "The steps performed include:\n",
+        "\n",
+        "- Create a Vertex `Dataset` resource.\n",
+        "- Train the model.\n",
+        "- View the model evaluation.\n",
+        "- Make a batch prediction.\n",
+        "\n",
+        "There is one key difference between using batch prediction and using online prediction:\n",
+        "\n",
+        "* Prediction Service: Does an on-demand prediction for the entire set of instances (i.e., one or more data items) and returns the results in real-time.\n",
+        "\n",
+        "* Batch Prediction Service: Does a queued (batch) prediction for the entire set of instances in the background and stores the results in a Cloud Storage bucket when ready."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "dataset:biomedical,ten"
+      },
+      "source": [
+        "### Dataset\n",
+        "\n",
+        "The dataset used for this tutorial is the [NCBI Disease Research Abstracts dataset](https://www.ncbi.nlm.nih.gov/CBBresearch/Dogan/DISEASE/) from [National Center for Biotechnology Information](https://www.ncbi.nlm.nih.gov/). The version of the dataset you will use in this tutorial is stored in a public Cloud Storage bucket."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "costs"
+      },
+      "source": [
+        "### Costs\n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI\n",
+        "* Cloud Storage\n",
+        "\n",
+        "Learn about [Vertex AI\n",
+        "pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage\n",
+        "pricing](https://cloud.google.com/storage/pricing), and use the [Pricing\n",
+        "Calculator](https://cloud.google.com/products/calculator/)\n",
+        "to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "db52a0a61fca"
+      },
+      "source": [
+        "### Installation\n",
+        "\n",
+        "Install the following packages for executing this notebook."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b75757581291"
+      },
+      "outputs": [],
+      "source": [
+        "# install packages\n",
+        "! pip3 install --upgrade --quiet google-cloud-aiplatform \\\n",
+        "                                    jsonlines -q\n",
+        "! pip3 install --upgrade tensorflow -q"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "e9255e3b156f"
+      },
+      "source": [
+        "### Colab Only: Uncomment the following cell to restart the kernel"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "0c0b2427998a"
+      },
+      "outputs": [],
+      "source": [
+        "# Automatically restart kernel after installs so that your environment can access the new packages\n",
+        "# import IPython\n",
+        "\n",
+        "# app = IPython.Application.instance()\n",
+        "# app.kernel.do_shutdown(True)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "435b8e413535"
+      },
+      "source": [
+        "### Before you begin\n",
+        "\n",
+        "#### Set your project ID\n",
+        "\n",
+        "**If you don't know your project ID**, try the following:\n",
+        "-  Run `gcloud config list`\n",
+        "-  Run `gcloud projects list`\n",
+        "-  See the support page: [Locate the project ID](https://support.google.com/googleapi/answer/7014113)"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "be175254a715"
+      },
+      "outputs": [],
+      "source": [
+        "PROJECT_ID = \"[your-project-id]\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# set the project id\n",
+        "! gcloud config set project $PROJECT_ID"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "2e6b8b324ce1"
+      },
+      "source": [
+        "#### Region\n",
+        "\n",
+        "You can also change the `REGION` variable used by Vertex AI. \n",
+        "Learn more about [Vertex AI regions](https://cloud.google.com/vertex-ai/docs/general/locations)."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "region"
+      },
+      "outputs": [],
+      "source": [
+        "REGION = \"us-central1\"  # @param {type: \"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "6c43a8673066"
+      },
+      "source": [
+        "### Authenticate your Google Cloud account\n",
+        "\n",
+        "Depending on your Jupyter environment, you may have to manually authenticate. Follow the relevant instructions below.\n",
+        "\n",
+        "**1. Vertex AI Workbench** \n",
+        "- Do nothing as you are already authenticated.\n",
+        "\n",
+        "**2. Local JupyterLab Instance,** uncomment and run."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "fbc9cd30cc4b"
+      },
+      "outputs": [],
+      "source": [
+        "# ! gcloud auth login"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "cd0da2c26879"
+      },
+      "source": [
+        "**3. Colab,** uncomment and run:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "a336a05c6149"
+      },
+      "outputs": [],
+      "source": [
+        "# from google.colab import auth\n",
+        "# auth.authenticate_user()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0461097edfa5"
+      },
+      "source": [
+        "**4. Service Account or other**\n",
+        "- See all the authentication options here: [Google Cloud Platform Jupyter Notebook Authentication Guide](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/notebook_authentication_guide.ipynb)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "e5755d1a554f"
+      },
+      "source": [
+        "### Create a Cloud Storage bucket\n",
+        "\n",
+        "Create a storage bucket to store intermediate artifacts such as datasets."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "d2de92accb67"
+      },
+      "outputs": [],
+      "source": [
+        "BUCKET_URI = f\"gs://your-bucket-name-unique-{PROJECT_ID}\"  # @param {type:\"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "b72bfdf29dae"
+      },
+      "source": [
+        "**Only if your bucket doesn't already exist**: Run the following cell to create your Cloud Storage bucket."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "create_bucket"
+      },
+      "outputs": [],
+      "source": [
+        "! gsutil mb -l $REGION $BUCKET_URI"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "setup_vars"
+      },
+      "source": [
+        "### Set up variables\n",
+        "\n",
+        "Next, set up some variables used throughout the tutorial.\n",
+        "### Import libraries and define constants"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "import_aip:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "import google.cloud.aiplatform as aiplatform"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "init_aip:mbsdk"
+      },
+      "source": [
+        "## Initialize Vertex AI SDK for Python\n",
+        "\n",
+        "Initialize the Vertex AI SDK for Python for your project and corresponding bucket."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "init_aip:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "aiplatform.init(project=PROJECT_ID, staging_bucket=BUCKET_URI)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "tutorial_start:automl"
+      },
+      "source": [
+        "# Tutorial\n",
+        "\n",
+        "Now you are ready to start creating your own AutoML text entity extraction model."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "import_file:u_dataset,jsonl"
+      },
+      "source": [
+        "#### Location of Cloud Storage training data.\n",
+        "\n",
+        "Now set the variable `IMPORT_FILE` to the location of the JSONL index file in Cloud Storage."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "import_file:biomedical,jsonl,ten"
+      },
+      "outputs": [],
+      "source": [
+        "IMPORT_FILE = \"gs://cloud-samples-data/language/ucaip_ten_dataset.jsonl\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "quick_peek:jsonl"
+      },
+      "source": [
+        "#### Quick peek at your data\n",
+        "\n",
+        "This tutorial uses a version of the NCBI Biomedical dataset that is stored in a public Cloud Storage bucket, using a JSONL index file.\n",
+        "\n",
+        "Start by doing a quick peek at the data. You count the number of examples by counting the number of objects in a JSONL index file  (`wc -l`) and then peek at the first few rows."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "quick_peek:jsonl"
+      },
+      "outputs": [],
+      "source": [
+        "if \"IMPORT_FILES\" in globals():\n",
+        "    FILE = IMPORT_FILES[0]\n",
+        "else:\n",
+        "    FILE = IMPORT_FILE\n",
+        "\n",
+        "count = ! gsutil cat $FILE | wc -l\n",
+        "print(\"Number of Examples\", int(count[0]))\n",
+        "\n",
+        "print(\"First 10 rows\")\n",
+        "! gsutil cat $FILE | head"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "create_dataset:text,ten"
+      },
+      "source": [
+        "### Create the Dataset\n",
+        "\n",
+        "Next, create the `Dataset` resource using the `create` method for the `TextDataset` class, which takes the following parameters:\n",
+        "\n",
+        "- `display_name`: The human readable name for the `Dataset` resource.\n",
+        "- `gcs_source`: A list of one or more dataset index files to import the data items into the `Dataset` resource.\n",
+        "- `import_schema_uri`: The data labeling schema for the data items.\n",
+        "\n",
+        "This operation may take several minutes."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "create_dataset:text,ten"
+      },
+      "outputs": [],
+      "source": [
+        "dataset = aiplatform.TextDataset.create(\n",
+        "    display_name=\"NCBI Biomedical\",\n",
+        "    gcs_source=[IMPORT_FILE],\n",
+        "    import_schema_uri=aiplatform.schema.dataset.ioformat.text.extraction,\n",
+        ")\n",
+        "\n",
+        "print(dataset.resource_name)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "create_automl_pipeline:text,ten"
+      },
+      "source": [
+        "### Create and run training pipeline\n",
+        "\n",
+        "To train an AutoML model, you perform two steps: 1) create a training pipeline, and 2) run the pipeline.\n",
+        "\n",
+        "#### Create training pipeline\n",
+        "\n",
+        "An AutoML training pipeline is created with the `AutoMLTextTrainingJob` class, with the following parameters:\n",
+        "\n",
+        "- `display_name`: The human readable name for the `TrainingJob` resource.\n",
+        "- `prediction_type`: The type task to train the model for.\n",
+        "  - `classification`: A text classification model.\n",
+        "  - `sentiment`: A text sentiment analysis model.\n",
+        "  - `extraction`: A text entity extraction model.\n",
+        "- `multi_label`: If a classification task, whether single (False) or multi-labeled (True).\n",
+        "- `sentiment_max`: If a sentiment analysis task, the maximum sentiment value.\n",
+        "\n",
+        "The instantiated object is the DAG (directed acyclic graph) for the training pipeline."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "create_automl_pipeline:text,ten"
+      },
+      "outputs": [],
+      "source": [
+        "dag = aiplatform.AutoMLTextTrainingJob(\n",
+        "    display_name=\"biomedical\", prediction_type=\"extraction\"\n",
+        ")\n",
+        "\n",
+        "print(dag)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "run_automl_pipeline:text"
+      },
+      "source": [
+        "#### Run the training pipeline\n",
+        "\n",
+        "Next, you run the DAG to start the training job by invoking the method `run`, with the following parameters:\n",
+        "\n",
+        "- `dataset`: The `Dataset` resource to train the model.\n",
+        "- `model_display_name`: The human readable name for the trained model.\n",
+        "- `training_fraction_split`: The percentage of the dataset to use for training.\n",
+        "- `test_fraction_split`: The percentage of the dataset to use for test (holdout data).\n",
+        "- `validation_fraction_split`: The percentage of the dataset to use for validation.\n",
+        "\n",
+        "The `run` method when completed returns the `Model` resource.\n",
+        "\n",
+        "The execution of the training pipeline will take upto 20 minutes."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "run_automl_pipeline:text"
+      },
+      "outputs": [],
+      "source": [
+        "model = dag.run(\n",
+        "    dataset=dataset,\n",
+        "    model_display_name=\"biomedical\",\n",
+        "    training_fraction_split=0.8,\n",
+        "    validation_fraction_split=0.1,\n",
+        "    test_fraction_split=0.1,\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "caaa3f32b12e"
+      },
+      "source": [
+        "## Review model evaluation scores\n",
+        "\n",
+        "After your model training has finished, you can review the evaluation scores for it using the `list_model_evaluations()` method. This method will return an iterator for each evaluation slice."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b0bb6be8621a"
+      },
+      "outputs": [],
+      "source": [
+        "model_evaluations = model.list_model_evaluations()\n",
+        "\n",
+        "for model_evaluation in model_evaluations:\n",
+        "    print(model_evaluation.to_dict())"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "make_prediction"
+      },
+      "source": [
+        "## Send a batch prediction request\n",
+        "\n",
+        "Send a batch prediction to deployed model."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "make_test_items:automl,batch_prediction"
+      },
+      "source": [
+        "### Make test items\n",
+        "\n",
+        "You use synthetic data as a test data items. Don't be concerned that we are using synthetic data -- we just want to demonstrate how to make a prediction."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "make_test_items:automl,text,biomedical"
+      },
+      "outputs": [],
+      "source": [
+        "test_item_1 = 'Molecular basis of hexosaminidase A deficiency and pseudodeficiency in the Berks County Pennsylvania Dutch.\\tFollowing the birth of two infants with Tay-Sachs disease ( TSD ) , a non-Jewish , Pennsylvania Dutch kindred was screened for TSD carriers using the biochemical assay . A high frequency of individuals who appeared to be TSD heterozygotes was detected ( Kelly et al . , 1975 ) . Clinical and biochemical evidence suggested that the increased carrier frequency was due to at least two altered alleles for the hexosaminidase A alpha-subunit . We now report two mutant alleles in this Pennsylvania Dutch kindred , and one polymorphism . One allele , reported originally in a French TSD patient ( Akli et al . , 1991 ) , is a GT-- > AT transition at the donor splice-site of intron 9 . The second , a C-- > T transition at nucleotide 739 ( Arg247Trp ) , has been shown by Triggs-Raine et al . ( 1992 ) to be a clinically benign \" pseudodeficient \" allele associated with reduced enzyme activity against artificial substrate . Finally , a polymorphism [ G-- > A ( 759 ) ] , which leaves valine at codon 253 unchanged , is described'\n",
+        "test_item_2 = \"Analysis of alkaptonuria (AKU) mutations and polymorphisms reveals that the CCC sequence motif is a mutational hot spot in the homogentisate 1,2 dioxygenase gene (HGO).\tWe recently showed that alkaptonuria ( AKU ) is caused by loss-of-function mutations in the homogentisate 1 , 2 dioxygenase gene ( HGO ) . Herein we describe haplotype and mutational analyses of HGO in seven new AKU pedigrees . These analyses identified two novel single-nucleotide polymorphisms ( INV4 + 31A-- > G and INV11 + 18A-- > G ) and six novel AKU mutations ( INV1-1G-- > A , W60G , Y62C , A122D , P230T , and D291E ) , which further illustrates the remarkable allelic heterogeneity found in AKU . Reexamination of all 29 mutations and polymorphisms thus far described in HGO shows that these nucleotide changes are not randomly distributed ; the CCC sequence motif and its inverted complement , GGG , are preferentially mutated . These analyses also demonstrated that the nucleotide substitutions in HGO do not involve CpG dinucleotides , which illustrates important differences between HGO and other genes for the occurrence of mutation at specific short-sequence motifs . Because the CCC sequence motifs comprise a significant proportion ( 34 . 5 % ) of all mutated bases that have been observed in HGO , we conclude that the CCC triplet is a mutational hot spot in HGO .\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "make_batch_file:automl,text"
+      },
+      "source": [
+        "### Make the batch input file\n",
+        "\n",
+        "Now make a batch input file, which you will store in your local Cloud Storage bucket. The batch input file can only be in JSONL format. For JSONL file, you make one dictionary entry per line for each data item (instance). The dictionary contains the key/value pairs:\n",
+        "\n",
+        "- `content`: The Cloud Storage path to the file with the text item.\n",
+        "- `mime_type`: The content type. In our example, it is a `text` file.\n",
+        "\n",
+        "For example:\n",
+        "\n",
+        "                        {'content': '[your-bucket]/file1.txt', 'mime_type': 'text'}"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "make_batch_file:automl,text"
+      },
+      "outputs": [],
+      "source": [
+        "import json\n",
+        "\n",
+        "import tensorflow as tf\n",
+        "\n",
+        "gcs_test_item_1 = BUCKET_URI + \"/test1.txt\"\n",
+        "with tf.io.gfile.GFile(gcs_test_item_1, \"w\") as f:\n",
+        "    f.write(test_item_1 + \"\\n\")\n",
+        "gcs_test_item_2 = BUCKET_URI + \"/test2.txt\"\n",
+        "with tf.io.gfile.GFile(gcs_test_item_2, \"w\") as f:\n",
+        "    f.write(test_item_2 + \"\\n\")\n",
+        "\n",
+        "gcs_input_uri = BUCKET_URI + \"/test.jsonl\"\n",
+        "with tf.io.gfile.GFile(gcs_input_uri, \"w\") as f:\n",
+        "    data = {\"content\": gcs_test_item_1, \"mime_type\": \"text/plain\"}\n",
+        "    f.write(json.dumps(data) + \"\\n\")\n",
+        "    data = {\"content\": gcs_test_item_2, \"mime_type\": \"text/plain\"}\n",
+        "    f.write(json.dumps(data) + \"\\n\")\n",
+        "\n",
+        "print(gcs_input_uri)\n",
+        "! gsutil cat $gcs_input_uri"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "batch_request:mbsdk"
+      },
+      "source": [
+        "### Make the batch prediction request\n",
+        "\n",
+        "Now that your Model resource is trained, you can make a batch prediction by invoking the batch_predict() method, with the following parameters:\n",
+        "\n",
+        "- `job_display_name`: The human readable name for the batch prediction job.\n",
+        "- `gcs_source`: A list of one or more batch request input files.\n",
+        "- `gcs_destination_prefix`: The Cloud Storage location for storing the batch prediction resuls.\n",
+        "- `sync`: If set to True, the call will block while waiting for the asynchronous batch job to complete."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "batch_request:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "batch_predict_job = model.batch_predict(\n",
+        "    job_display_name=\"biomedical\",\n",
+        "    gcs_source=gcs_input_uri,\n",
+        "    gcs_destination_prefix=BUCKET_URI,\n",
+        "    sync=False,\n",
+        ")\n",
+        "\n",
+        "print(batch_predict_job)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "batch_request_wait:mbsdk"
+      },
+      "source": [
+        "### Wait for completion of batch prediction job\n",
+        "\n",
+        "Next, wait for the batch job to complete. Alternatively, one can set the parameter `sync` to `True` in the `batch_predict()` method to block until the batch prediction job is completed."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "batch_request_wait:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "batch_predict_job.wait()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "get_batch_prediction:mbsdk,ten"
+      },
+      "source": [
+        "### Get the predictions\n",
+        "\n",
+        "Next, get the results from the completed batch prediction job.\n",
+        "\n",
+        "The results are written to the Cloud Storage output bucket you specified in the batch prediction request. You call the method iter_outputs() to get a list of each Cloud Storage file generated with the results. Each file contains one or more prediction requests in a JSON format:\n",
+        "\n",
+        "- `content`: The prediction request.\n",
+        "- `prediction`: The prediction response.\n",
+        " - `ids`: The internal assigned unique identifiers for each prediction request.\n",
+        " - `displayNames`: The class names for each class label.\n",
+        " - `confidences`: The predicted confidence, between 0 and 1, per class label.\n",
+        " - `textSegmentStartOffsets`: The character offset in the text to the start of the entity.\n",
+        " - `textSegmentEndOffsets`: The character offset in the text to the end of the entity."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "get_batch_prediction:mbsdk,ten"
+      },
+      "outputs": [],
+      "source": [
+        "import json\n",
+        "\n",
+        "import tensorflow as tf\n",
+        "\n",
+        "bp_iter_outputs = batch_predict_job.iter_outputs()\n",
+        "\n",
+        "prediction_results = list()\n",
+        "for blob in bp_iter_outputs:\n",
+        "    if blob.name.split(\"/\")[-1].startswith(\"prediction\"):\n",
+        "        prediction_results.append(blob.name)\n",
+        "\n",
+        "tags = list()\n",
+        "for prediction_result in prediction_results:\n",
+        "    gfile_name = f\"gs://{bp_iter_outputs.bucket.name}/{prediction_result}\"\n",
+        "    with tf.io.gfile.GFile(name=gfile_name, mode=\"r\") as gfile:\n",
+        "        for line in gfile.readlines():\n",
+        "            line = json.loads(line)\n",
+        "            print(line)\n",
+        "            break"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "cleanup:mbsdk"
+      },
+      "source": [
+        "# Cleaning up\n",
+        "\n",
+        "To clean up all Google Cloud resources used in this project, you can [delete the Google Cloud\n",
+        "project](https://cloud.google.com/resource-manager/docs/creating-managing-projects#shutting_down_projects) you used for the tutorial.\n",
+        "\n",
+        "Otherwise, you can delete the individual resources you created in this tutorial."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "cleanup:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "import os\n",
+        "\n",
+        "delete_bucket = False\n",
+        "\n",
+        "if delete_bucket or os.getenv(\"IS_TESTING\"):\n",
+        "    ! gsutil rm -r $BUCKET_URI\n",
+        "\n",
+        "# Delete batch\n",
+        "batch_predict_job.delete()\n",
+        "\n",
+        "# Delete model\n",
+        "model.delete()\n",
+        "\n",
+        "# Delete text dataset\n",
+        "dataset.delete()\n",
+        "\n",
+        "# Delete training job\n",
+        "job.delete()"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "name": "automl_text_entity_extraction_batch_prediction.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/official/automl/automl_text_sentiment_analysis_batch_prediction.ipynb

@@ -0,0 +1,824 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "copyright"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2021 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "title"
+      },
+      "source": [
+        "# AutoML training text sentiment analysis model for batch prediction\n",
+        "\n",
+        "<table align=\"left\">\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_text_sentiment_analysis_batch_prediction.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_text_sentiment_analysis_batch_prediction.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/workbench/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/main/notebooks/official/automl/automl_text_sentiment_analysis_batch_prediction.ipynb\">\n",
+        "        <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "      Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>\n",
+        "<br/><br/><br/>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "overview:automl"
+      },
+      "source": [
+        "## Overview\n",
+        "\n",
+        "\n",
+        "This tutorial demonstrates how to use the Vertex AI SDK to create text sentiment analysis models and do batch prediction using a Google Cloud [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) model."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "objective:automl,training,batch_prediction"
+      },
+      "source": [
+        "### Objective\n",
+        "\n",
+        "In this tutorial, you create an AutoML text sentiment analysis model from a Python script, and then do a batch prediction using the Vertex SDK. You can alternatively create and deploy models using the `gcloud` command-line tool or online using the Cloud Console.\n",
+        "\n",
+        "The steps performed include:\n",
+        "\n",
+        "- Create a Vertex `Dataset` resource.\n",
+        "- Train the model.\n",
+        "- View the model evaluation.\n",
+        "- Make a batch prediction.\n",
+        "\n",
+        "There is one key difference between using batch prediction and using online prediction:\n",
+        "\n",
+        "* Prediction Service: Does an on-demand prediction for the entire set of instances (i.e., one or more data items) and returns the results in real-time.\n",
+        "\n",
+        "* Batch Prediction Service: Does a queued (batch) prediction for the entire set of instances in the background and stores the results in a Cloud Storage bucket when ready."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "dataset:claritin,tst"
+      },
+      "source": [
+        "### Dataset\n",
+        "\n",
+        "The dataset used for this tutorial is the [Crowdflower Claritin-Twitter dataset](https://data.world/crowdflower/claritin-twitter) from [data.world Datasets](https://data.world). The version of the dataset you will use in this tutorial is stored in a public Cloud Storage bucket."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "costs"
+      },
+      "source": [
+        "### Costs\n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI\n",
+        "* Cloud Storage\n",
+        "\n",
+        "Learn about [Vertex AI\n",
+        "pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage\n",
+        "pricing](https://cloud.google.com/storage/pricing), and use the [Pricing\n",
+        "Calculator](https://cloud.google.com/products/calculator/)\n",
+        "to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "db52a0a61fca"
+      },
+      "source": [
+        "### Installation\n",
+        "\n",
+        "Install the following packages for executing this notebook."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "install_aip:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "! pip3 install --upgrade google-cloud-aiplatform -q\n",
+        "\n",
+        "! pip3 install --upgrade tensorflow -q"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "e9255e3b156f"
+      },
+      "source": [
+        "### Colab Only: Uncomment the following cell to restart the kernel"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "0c0b2427998a"
+      },
+      "outputs": [],
+      "source": [
+        "# Automatically restart kernel after installs so that your environment can access the new packages\n",
+        "# import IPython\n",
+        "\n",
+        "# app = IPython.Application.instance()\n",
+        "# app.kernel.do_shutdown(True)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "435b8e413535"
+      },
+      "source": [
+        "### Before you begin\n",
+        "\n",
+        "#### Set your project ID\n",
+        "\n",
+        "**If you don't know your project ID**, try the following:\n",
+        "-  Run `gcloud config list`\n",
+        "-  Run `gcloud projects list`\n",
+        "-  See the support page: [Locate the project ID](https://support.google.com/googleapi/answer/7014113)"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "set_project_id"
+      },
+      "outputs": [],
+      "source": [
+        "PROJECT_ID = \"[your-project-id]\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# set the project id\n",
+        "! gcloud config set project $PROJECT_ID"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "2e6b8b324ce1"
+      },
+      "source": [
+        "#### Region\n",
+        "\n",
+        "You can also change the `REGION` variable used by Vertex AI. \n",
+        "Learn more about [Vertex AI regions](https://cloud.google.com/vertex-ai/docs/general/locations)."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "region"
+      },
+      "outputs": [],
+      "source": [
+        "REGION = \"us-central1\"  # @param {type: \"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "6c43a8673066"
+      },
+      "source": [
+        "### Authenticate your Google Cloud account\n",
+        "\n",
+        "Depending on your Jupyter environment, you may have to manually authenticate. Follow the relevant instructions below.\n",
+        "\n",
+        "**1. Vertex AI Workbench** \n",
+        "- Do nothing as you are already authenticated.\n",
+        "\n",
+        "**2. Local JupyterLab Instance,** uncomment and run."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "fbc9cd30cc4b"
+      },
+      "outputs": [],
+      "source": [
+        "# ! gcloud auth login"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "cd0da2c26879"
+      },
+      "source": [
+        "**3. Colab,** uncomment and run:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "a336a05c6149"
+      },
+      "outputs": [],
+      "source": [
+        "# from google.colab import auth\n",
+        "# auth.authenticate_user()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0461097edfa5"
+      },
+      "source": [
+        "**4. Service Account or other**\n",
+        "- See all the authentication options here: [Google Cloud Platform Jupyter Notebook Authentication Guide](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/notebook_authentication_guide.ipynb)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "e5755d1a554f"
+      },
+      "source": [
+        "### Create a Cloud Storage bucket\n",
+        "\n",
+        "Create a storage bucket to store intermediate artifacts such as datasets."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "d2de92accb67"
+      },
+      "outputs": [],
+      "source": [
+        "BUCKET_URI = f\"gs://your-bucket-name-unique-{PROJECT_ID}\"  # @param {type:\"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "create_bucket"
+      },
+      "source": [
+        "**Only if your bucket doesn't already exist**: Run the following cell to create your Cloud Storage bucket."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "create_bucket"
+      },
+      "outputs": [],
+      "source": [
+        "! gsutil mb -l $REGION $BUCKET_URI"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "setup_vars"
+      },
+      "source": [
+        "### Set up variables\n",
+        "\n",
+        "Next, set up some variables used throughout the tutorial.\n",
+        "### Import libraries and define constants"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "import_aip:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "import google.cloud.aiplatform as aiplatform"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "init_aip:mbsdk"
+      },
+      "source": [
+        "## Initialize Vertex AI SDK for Python\n",
+        "\n",
+        "Initialize the Vertex AI SDK for Python for your project and corresponding bucket."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "init_aip:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "aiplatform.init(project=PROJECT_ID, staging_bucket=BUCKET_URI)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "tutorial_start:automl"
+      },
+      "source": [
+        "# Tutorial\n",
+        "\n",
+        "Now you are ready to start creating your own AutoML text sentiment analysis model."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "import_file:u_dataset,csv"
+      },
+      "source": [
+        "#### Location of Cloud Storage training data.\n",
+        "\n",
+        "Now set the variable `IMPORT_FILE` to the location of the CSV index file in Cloud Storage."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "import_file:claritin,csv,tst"
+      },
+      "outputs": [],
+      "source": [
+        "IMPORT_FILE = \"gs://cloud-samples-data/language/claritin.csv\"\n",
+        "SENTIMENT_MAX = 4"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "quick_peek:csv"
+      },
+      "source": [
+        "#### Quick peek at your data\n",
+        "\n",
+        "This tutorial uses a version of the Crowdflower Claritin-Twitter dataset that is stored in a public Cloud Storage bucket, using a CSV index file.\n",
+        "\n",
+        "Start by doing a quick peek at the data. You count the number of examples by counting the number of rows in the CSV index file  (`wc -l`) and then peek at the first few rows."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "quick_peek:csv"
+      },
+      "outputs": [],
+      "source": [
+        "if \"IMPORT_FILES\" in globals():\n",
+        "    FILE = IMPORT_FILES[0]\n",
+        "else:\n",
+        "    FILE = IMPORT_FILE\n",
+        "\n",
+        "count = ! gsutil cat $FILE | wc -l\n",
+        "print(\"Number of Examples\", int(count[0]))\n",
+        "\n",
+        "print(\"First 10 rows\")\n",
+        "! gsutil cat $FILE | head"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "create_dataset:text,tst"
+      },
+      "source": [
+        "### Create the Dataset\n",
+        "\n",
+        "Next, create the `Dataset` resource using the `create` method for the `TextDataset` class, which takes the following parameters:\n",
+        "\n",
+        "- `display_name`: The human readable name for the `Dataset` resource.\n",
+        "- `gcs_source`: A list of one or more dataset index files to import the data items into the `Dataset` resource.\n",
+        "- `import_schema_uri`: The data labeling schema for the data items.\n",
+        "\n",
+        "This operation may take several minutes."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "create_dataset:text,tst"
+      },
+      "outputs": [],
+      "source": [
+        "dataset = aiplatform.TextDataset.create(\n",
+        "    display_name=\"Crowdflower Claritin-Twitter\",\n",
+        "    gcs_source=[IMPORT_FILE],\n",
+        "    import_schema_uri=aiplatform.schema.dataset.ioformat.text.sentiment,\n",
+        ")\n",
+        "\n",
+        "print(dataset.resource_name)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "create_automl_pipeline:text,tst"
+      },
+      "source": [
+        "### Create and run training pipeline\n",
+        "\n",
+        "To train an AutoML model, you perform two steps: 1) create a training pipeline, and 2) run the pipeline.\n",
+        "\n",
+        "#### Create training pipeline\n",
+        "\n",
+        "An AutoML training pipeline is created with the `AutoMLTextTrainingJob` class, with the following parameters:\n",
+        "\n",
+        "- `display_name`: The human readable name for the `TrainingJob` resource.\n",
+        "- `prediction_type`: The type task to train the model for.\n",
+        "  - `classification`: A text classification model.\n",
+        "  - `sentiment`: A text sentiment analysis model.\n",
+        "  - `extraction`: A text entity extraction model.\n",
+        "- `multi_label`: If a classification task, whether single (False) or multi-labeled (True).\n",
+        "- `sentiment_max`: If a sentiment analysis task, the maximum sentiment value.\n",
+        "\n",
+        "The instantiated object is the DAG (directed acyclic graph) for the training pipeline."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "create_automl_pipeline:text,tst"
+      },
+      "outputs": [],
+      "source": [
+        "dag = aiplatform.AutoMLTextTrainingJob(\n",
+        "    display_name=\"claritin\",\n",
+        "    prediction_type=\"sentiment\",\n",
+        "    sentiment_max=SENTIMENT_MAX,\n",
+        ")\n",
+        "\n",
+        "print(dag)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "run_automl_pipeline:text"
+      },
+      "source": [
+        "#### Run the training pipeline\n",
+        "\n",
+        "Next, you run the DAG to start the training job by invoking the method `run`, with the following parameters:\n",
+        "\n",
+        "- `dataset`: The `Dataset` resource to train the model.\n",
+        "- `model_display_name`: The human readable name for the trained model.\n",
+        "- `training_fraction_split`: The percentage of the dataset to use for training.\n",
+        "- `test_fraction_split`: The percentage of the dataset to use for test (holdout data).\n",
+        "- `validation_fraction_split`: The percentage of the dataset to use for validation.\n",
+        "\n",
+        "The `run` method when completed returns the `Model` resource.\n",
+        "\n",
+        "The execution of the training pipeline will take upto 20 minutes."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "run_automl_pipeline:text"
+      },
+      "outputs": [],
+      "source": [
+        "model = dag.run(\n",
+        "    dataset=dataset,\n",
+        "    model_display_name=\"claritin\",\n",
+        "    training_fraction_split=0.8,\n",
+        "    validation_fraction_split=0.1,\n",
+        "    test_fraction_split=0.1,\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "caaa3f32b12e"
+      },
+      "source": [
+        "## Review model evaluation scores\n",
+        "\n",
+        "After your model training has finished, you can review the evaluation scores for it using the `list_model_evaluations()` method. This method will return an iterator for each evaluation slice."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "b0bb6be8621a"
+      },
+      "outputs": [],
+      "source": [
+        "model_evaluations = model.list_model_evaluations()\n",
+        "\n",
+        "for model_evaluation in model_evaluations:\n",
+        "    print(model_evaluation.to_dict())"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "make_prediction"
+      },
+      "source": [
+        "## Send a batch prediction request\n",
+        "\n",
+        "Send a batch prediction to your model."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "get_test_items:batch_prediction"
+      },
+      "source": [
+        "### Get test item(s)\n",
+        "\n",
+        "Now do a batch prediction to your Vertex model. You will use arbitrary examples out of the dataset as a test items. Don't be concerned that the examples were likely used in training the model -- we just want to demonstrate how to make a prediction."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "get_test_items:automl,tst,csv"
+      },
+      "outputs": [],
+      "source": [
+        "test_items = ! gsutil cat $IMPORT_FILE | head -n2\n",
+        "\n",
+        "if len(test_items[0]) == 4:\n",
+        "    _, test_item_1, test_label_1, _ = str(test_items[0]).split(\",\")\n",
+        "    _, test_item_2, test_label_2, _ = str(test_items[1]).split(\",\")\n",
+        "else:\n",
+        "    test_item_1, test_label_1, _ = str(test_items[0]).split(\",\")\n",
+        "    test_item_2, test_label_2, _ = str(test_items[1]).split(\",\")\n",
+        "\n",
+        "\n",
+        "print(test_item_1, test_label_1)\n",
+        "print(test_item_2, test_label_2)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "make_batch_file:automl,text"
+      },
+      "source": [
+        "### Make the batch input file\n",
+        "\n",
+        "Now make a batch input file, which you will store in your local Cloud Storage bucket. The batch input file can only be in JSONL format. For JSONL file, you make one dictionary entry per line for each data item (instance). The dictionary contains the key/value pairs:\n",
+        "\n",
+        "- `content`: The Cloud Storage path to the file with the text item.\n",
+        "- `mime_type`: The content type. In our example, it is a `text` file.\n",
+        "\n",
+        "For example:\n",
+        "\n",
+        "                        {'content': '[your-bucket]/file1.txt', 'mime_type': 'text'}"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "make_batch_file:automl,text"
+      },
+      "outputs": [],
+      "source": [
+        "import json\n",
+        "\n",
+        "import tensorflow as tf\n",
+        "\n",
+        "gcs_test_item_1 = BUCKET_URI + \"/test1.txt\"\n",
+        "with tf.io.gfile.GFile(gcs_test_item_1, \"w\") as f:\n",
+        "    f.write(test_item_1 + \"\\n\")\n",
+        "gcs_test_item_2 = BUCKET_URI + \"/test2.txt\"\n",
+        "with tf.io.gfile.GFile(gcs_test_item_2, \"w\") as f:\n",
+        "    f.write(test_item_2 + \"\\n\")\n",
+        "\n",
+        "gcs_input_uri = BUCKET_URI + \"/test.jsonl\"\n",
+        "with tf.io.gfile.GFile(gcs_input_uri, \"w\") as f:\n",
+        "    data = {\"content\": gcs_test_item_1, \"mime_type\": \"text/plain\"}\n",
+        "    f.write(json.dumps(data) + \"\\n\")\n",
+        "    data = {\"content\": gcs_test_item_2, \"mime_type\": \"text/plain\"}\n",
+        "    f.write(json.dumps(data) + \"\\n\")\n",
+        "\n",
+        "print(gcs_input_uri)\n",
+        "! gsutil cat $gcs_input_uri"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "batch_request:mbsdk"
+      },
+      "source": [
+        "### Make the batch prediction request\n",
+        "\n",
+        "Now that your Model resource is trained, you can make a batch prediction by invoking the batch_predict() method, with the following parameters:\n",
+        "\n",
+        "- `job_display_name`: The human readable name for the batch prediction job.\n",
+        "- `gcs_source`: A list of one or more batch request input files.\n",
+        "- `gcs_destination_prefix`: The Cloud Storage location for storing the batch prediction resuls.\n",
+        "- `sync`: If set to True, the call will block while waiting for the asynchronous batch job to complete."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "batch_request:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "batch_predict_job = model.batch_predict(\n",
+        "    job_display_name=\"claritin\",\n",
+        "    gcs_source=gcs_input_uri,\n",
+        "    gcs_destination_prefix=BUCKET_URI,\n",
+        "    sync=False,\n",
+        ")\n",
+        "\n",
+        "print(batch_predict_job)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "batch_request_wait:mbsdk"
+      },
+      "source": [
+        "### Wait for completion of batch prediction job\n",
+        "\n",
+        "Next, wait for the batch job to complete. Alternatively, one can set the parameter `sync` to `True` in the `batch_predict()` method to block until the batch prediction job is completed."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "batch_request_wait:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "batch_predict_job.wait()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "get_batch_prediction:mbsdk,tst"
+      },
+      "source": [
+        "### Get the predictions\n",
+        "\n",
+        "Next, get the results from the completed batch prediction job.\n",
+        "\n",
+        "The results are written to the Cloud Storage output bucket you specified in the batch prediction request. You call the method iter_outputs() to get a list of each Cloud Storage file generated with the results. Each file contains one or more prediction requests in a JSON format:\n",
+        "\n",
+        "- `content`: The prediction request.\n",
+        "- `prediction`: The prediction response.\n",
+        " - `sentiment`: The sentiment."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "get_batch_prediction:mbsdk,tst"
+      },
+      "outputs": [],
+      "source": [
+        "import json\n",
+        "\n",
+        "import tensorflow as tf\n",
+        "\n",
+        "bp_iter_outputs = batch_predict_job.iter_outputs()\n",
+        "\n",
+        "prediction_results = list()\n",
+        "for blob in bp_iter_outputs:\n",
+        "    if blob.name.split(\"/\")[-1].startswith(\"prediction\"):\n",
+        "        prediction_results.append(blob.name)\n",
+        "\n",
+        "tags = list()\n",
+        "for prediction_result in prediction_results:\n",
+        "    gfile_name = f\"gs://{bp_iter_outputs.bucket.name}/{prediction_result}\"\n",
+        "    with tf.io.gfile.GFile(name=gfile_name, mode=\"r\") as gfile:\n",
+        "        for line in gfile.readlines():\n",
+        "            line = json.loads(line)\n",
+        "            print(line)\n",
+        "            break"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "cleanup:mbsdk"
+      },
+      "source": [
+        "# Cleaning up\n",
+        "\n",
+        "To clean up all Google Cloud resources used in this project, you can [delete the Google Cloud\n",
+        "project](https://cloud.google.com/resource-manager/docs/creating-managing-projects#shutting_down_projects) you used for the tutorial.\n",
+        "\n",
+        "Otherwise, you can delete the individual resources you created in this tutorial."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "cleanup:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "import os\n",
+        "\n",
+        "delete_bucket = False\n",
+        "\n",
+        "if delete_bucket or os.getenv(\"IS_TESTING\"):\n",
+        "    ! gsutil rm -r $BUCKET_URI\n",
+        "\n",
+        "# Delete batch\n",
+        "batch_predict_job.delete()\n",
+        "\n",
+        "# Delete model\n",
+        "model.delete()\n",
+        "\n",
+        "# Delete text dataset\n",
+        "dataset.delete()\n",
+        "\n",
+        "# Delete training job\n",
+        "job.delete()"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "name": "automl_text_sentiment_analysis_batch_prediction.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/official/automl/sdk_automl_image_object_detection_batch.ipynb

@@ -63,7 +63,9 @@
         "## Overview\n",
         "\n",
         "\n",
-        "This tutorial demonstrates how to use the Vertex AI SDK to create image object detection models and do batch prediction using a Google Cloud [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) model."
+        "This tutorial demonstrates how to use the Vertex AI SDK to create image object detection models and do batch prediction using a Google Cloud [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) model.\n",
+        "\n",
+        "Learn more about [Object detection for image data](https://cloud.google.com/vertex-ai/docs/training-overview#object_detection_for_images)."
       ]
     },
     {

notebooks/official/automl/sdk_automl_tabular_forecasting_batch.ipynb

@@ -62,7 +62,9 @@
         "## Overview\n",
         "\n",
         "\n",
-        "This tutorial demonstrates how to use the Vertex AI SDK to create tabular forecasting models and do batch prediction using a Google Cloud [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) model."
+        "This tutorial demonstrates how to use the Vertex AI SDK to create tabular forecasting models and do batch prediction using a Google Cloud [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) model.\n",
+        "\n",
+        "Learn more about [Forecasting for tabular data](https://cloud.google.com/vertex-ai/docs/tabular-data/forecasting/overview)."
       ]
     },
     {
@@ -831,6 +833,7 @@
       ]
     },
     {
+      "attachments": {},
       "cell_type": "markdown",
       "metadata": {
         "id": "99b7a9287ba6"
@@ -838,7 +841,7 @@
       "source": [
         "For AutoML models, manual scaling can be adjusted by setting both min and max nodes i.e., `starting_replica_count` and `max_replica_count` as the same value(in this example, set to 1). The node count can be increased or decreased as required by load.\n",
         " \n",
-        "`batch_predict` can export predictions either to BigQuery or GCS. The BigQuery options are commented out below and the predictions will be exported to the BUCKET_URI."
+        "`batch_predict` can export predictions either to BigQuery or GCS. This example exports to BigQuery."
       ]
     },
     {

notebooks/official/automl/sdk_automl_tabular_regression_batch_bq.ipynb

@@ -61,7 +61,9 @@
         "## Overview\n",
         "\n",
         "\n",
-        "This tutorial demonstrates how to use the Vertex AI SDK for Python to create tabular regression models and do batch prediction using a Google Cloud [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) model."
+        "This tutorial demonstrates how to use the Vertex AI SDK for Python to create tabular regression models and do batch prediction using a Google Cloud [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) model.\n",
+        "\n",
+        "Learn more about [Regression for tabular data](https://cloud.google.com/vertex-ai/docs/tabular-data/classification-regression/overview)."
       ]
     },
     {

notebooks/official/automl/sdk_automl_tabular_regression_online_bq.ipynb

@@ -62,7 +62,9 @@
         "## Overview\n",
         "\n",
         "\n",
-        "This tutorial demonstrates how to use the Vertex AI SDK for Python to create tabular regression models and do online prediction using a Google Cloud [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) model."
+        "This tutorial demonstrates how to use the Vertex AI SDK for Python to create tabular regression models and do online prediction using a Google Cloud [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) model.\n",
+        "\n",
+        "Learn more about [Regression for tabular data](https://cloud.google.com/vertex-ai/docs/tabular-data/classification-regression/overview)."
       ]
     },
     {

notebooks/official/automl/sdk_automl_text_entity_extraction_online.ipynb

@@ -62,7 +62,9 @@
         "## Overview\n",
         "\n",
         "\n",
-        "This tutorial demonstrates how to use the Vertex AI SDK for Python to create text entity extraction models and do online prediction using a Google Cloud [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) model."
+        "This tutorial demonstrates how to use the Vertex AI SDK for Python to create text entity extraction models and do online prediction using a Google Cloud [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) model.\n",
+        "\n",
+        "Learn more about [Entity extraction for text data](https://cloud.google.com/vertex-ai/docs/training-overview#entity_extraction_for_text)."
       ]
     },
     {

notebooks/official/automl/sdk_automl_text_sentiment_analysis_online.ipynb

@@ -61,7 +61,9 @@
       "source": [
         "## Overview\n",
         "\n",
-        "This tutorial demonstrates how to use the Vertex AI SDK to train and deploy an [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) text sentiment analysis model and get online predictions from it."
+        "This tutorial demonstrates how to use the Vertex AI SDK to train and deploy an [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) text sentiment analysis model and get online predictions from it.\n",
+        "\n",
+        "Learn more about [Sentiment analysis for text data](https://cloud.google.com/vertex-ai/docs/training-overview#sentiment_analysis_for_text)."
       ]
     },
     {

notebooks/official/automl/sdk_automl_video_action_recognition_batch.ipynb

@@ -63,7 +63,9 @@
         "## Overview\n",
         "\n",
         "\n",
-        "This tutorial demonstrates how to use the Vertex AI SDK to create video action recognition models and do batch prediction using a Google Cloud [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) model."
+        "This tutorial demonstrates how to use the Vertex AI SDK to create video action recognition models and do batch prediction using a Google Cloud [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) model.\n",
+        "\n",
+        "Learn more about [Action recognition for video data](https://cloud.google.com/vertex-ai/docs/training-overview#action_recognition_for_videos)."
       ]
     },
     {

notebooks/official/automl/sdk_automl_video_classification_batch.ipynb

@@ -62,7 +62,9 @@
         "## Overview\n",
         "\n",
         "\n",
-        "This tutorial demonstrates how to use the Vertex AI SDK to create video classification models and do batch prediction using a Google Cloud [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) model."
+        "This tutorial demonstrates how to use the Vertex AI SDK to create video classification models and do batch prediction using a Google Cloud [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) model.\n",
+        "\n",
+        "Learn more about [Classification for video data](https://cloud.google.com/vertex-ai/docs/training-overview#classification_for_videos)."
       ]
     },
     {

notebooks/official/automl/sdk_automl_video_object_tracking_batch.ipynb

@@ -62,7 +62,9 @@
         "## Overview\n",
         "\n",
         "\n",
-        "This tutorial demonstrates how to use the Vertex AI SDK for Python to create video object tracking models and do batch prediction using a Google Cloud [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) model."
+        "This tutorial demonstrates how to use the Vertex AI SDK for Python to create video object tracking models and do batch prediction using a Google Cloud [AutoML](https://cloud.google.com/vertex-ai/docs/start/automl-users) model.\n",
+        "\n",
+        "Learn more about [Object tracking for video data](https://cloud.google.com/vertex-ai/docs/training-overview#object_tracking_for_videos)."
       ]
     },
     {

notebooks/official/bigquery_ml/README.md

@@ -1,3 +1,4 @@
+
 [Online prediction with BigQuery ML](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/bigquery_ml/bqml-online-prediction.ipynb)
 
 ```
@@ -14,3 +15,25 @@ The steps performed include:
 
 ```
 
+   Learn more about [BigQuery ML](https://cloud.google.com/vertex-ai/docs/beginner/bqml).
+
+
+[Get started with BigQuery ML Training](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/bigquery_ml/get_started_with_bqml_training.ipynb)
+
+```
+Learn how to use `BigQueryML` for training with `Vertex AI`.
+
+The steps performed include:
+
+- Create a local BigQuery table in your project
+- Train a BigQuery ML model
+- Evaluate the BigQuery ML model
+- Export the BigQuery ML model as a cloud model
+- Upload the exported model as a `Vertex AI Model` resource
+- Hyperparameter tune a BigQuery ML model with `Vertex AI Vizier`
+- Automatically register a BigQuery ML model to `Vertex AI Model Registry`
+
+```
+
+   Learn more about [BigQuery ML](https://cloud.google.com/vertex-ai/docs/beginner/bqml).
+

notebooks/official/bigquery_ml/bqml-online-prediction.ipynb

@@ -61,7 +61,9 @@
       "source": [
         "## Overview\n",
         "\n",
-        "This notebook is aimed at data analysts and data scientists who have data in BigQuery, want to train a model using BigQuery ML, register the model to Vertex AI Model Registry, and deploy it to an endpoint for real-time prediction. "
+        "This notebook is aimed at data analysts and data scientists who have data in BigQuery, want to train a model using BigQuery ML, register the model to Vertex AI Model Registry, and deploy it to an endpoint for real-time prediction. \n",
+        "\n",
+        "Learn more about [BigQuery ML](https://cloud.google.com/vertex-ai/docs/beginner/bqml)."
       ]
     },
     {

notebooks/official/custom/README.md

@@ -1,3 +1,24 @@
+
+[Deploying Iris-detection model using FastAPI and Vertex AI custom container serving](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/custom/SDK_Custom_Container_Prediction.ipynb)
+
+```
+Learn how to create, deploy and serve a custom classification model on Vertex AI.
+
+The steps performed include:
+
+- Train a model that uses flower's measurements as input to predict the class of iris.
+- Save the model and its serialized pre-processor.
+- Build a FastAPI server to handle predictions and health checks.
+- Build a custom container with model artifacts.
+- Upload and deploy custom container to Vertex AI Endpoints.
+
+```
+
+   Learn more about [Custom training](https://cloud.google.com/vertex-ai/docs/training/custom-training).
+
+   Learn more about [Vertex AI Prediction](https://cloud.google.com/vertex-ai/docs/predictions/get-predictions).
+
+
 [Training and deploying a sales forecasting model using FBProphet and Vertex AI](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/custom/SDK_FBProphet_Forecasting_Online.ipynb)
 
 ```
@@ -16,34 +37,9 @@ The steps performed include:
 
 ```
 
+   Learn more about [Custom training](https://cloud.google.com/vertex-ai/docs/training/custom-training).
 
-[Custom training and batch prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/custom/sdk-custom-image-classification-batch.ipynb)
-
-```
-Learn to use `Vertex AI Training` to create a custom trained model and use `Vertex AI Batch Prediction` to do a batch prediction on the trained model.
-
-The steps performed include:
-
-- Create a `Vertex AI` custom job for training a TensorFlow model.
-- Upload the trained model artifacts as a `Model` resource.
-- Make a batch prediction.
-
-```
-
-
-[Profile model training performance using Profiler](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/custom/custom_training_tensorboard_profiler.ipynb)
-
-```
-Learn how to enable Vertex AI TensorBoard Profiler for custom training jobs.
-
-The steps performed include:
-
-- Setup a service account and a Cloud Storage bucket
-- Create a TensorBoard instance
-- Create and run a custom training job
-- View the TensorBoard Profiler dashboard
-
-```
+   Learn more about [Vertex AI Prediction](https://cloud.google.com/vertex-ai/docs/predictions/get-predictions).
 
 
 [Training a TensorFlow model on BigQuery data](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/custom/custom-tabular-bq-managed-dataset.ipynb)
@@ -61,6 +57,43 @@ The steps performed include:
 
 ```
 
+   Learn more about [Vertex AI Training](https://cloud.google.com/vertex-ai/docs/training/custom-training).
+
+
+[Profile model training performance using Profiler](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/custom/custom_training_tensorboard_profiler.ipynb)
+
+```
+Learn how to enable Vertex AI TensorBoard Profiler for custom training jobs.
+
+The steps performed include:
+
+- Setup a service account and a Cloud Storage bucket
+- Create a TensorBoard instance
+- Create and run a custom training job
+- View the TensorBoard Profiler dashboard
+
+```
+
+   Learn more about [Vertex AI TensorBoard Profiler](https://cloud.google.com/vertex-ai/docs/experiments/tensorboard-profiler).
+
+
+[Custom training and batch prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/custom/sdk-custom-image-classification-batch.ipynb)
+
+```
+Learn to use `Vertex AI Training` to create a custom trained model and use `Vertex AI Batch Prediction` to do a batch prediction on the trained model.
+
+The steps performed include:
+
+- Create a `Vertex AI` custom job for training a TensorFlow model.
+- Upload the trained model artifacts as a `Model` resource.
+- Make a batch prediction.
+
+```
+
+   Learn more about [Custom training](https://cloud.google.com/vertex-ai/docs/training/custom-training).
+
+   Learn more about [Vertex AI Batch Prediction](https://cloud.google.com/vertex-ai/docs/tabular-data/classification-regression/get-batch-predictions).
+
 
 [Custom training and online prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/custom/sdk-custom-image-classification-online.ipynb)
 
@@ -78,19 +111,7 @@ The steps performed include:
 
 ```
 
+   Learn more about [Custom training](https://cloud.google.com/vertex-ai/docs/training/custom-training).
 
-[Deploying Iris-detection model using FastAPI and Vertex AI custom container serving](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/custom/SDK_Custom_Container_Prediction.ipynb)
-
-```
-Learn how to create, deploy and serve a custom classification model on Vertex AI.
-
-The steps performed include:
-
-- Train a model that uses flower's measurements as input to predict the class of iris.
-- Save the model and its serialized pre-processor.
-- Build a FastAPI server to handle predictions and health checks.
-- Build a custom container with model artifacts.
-- Upload and deploy custom container to Vertex AI Endpoints.
-
-```
+   Learn more about [Vertex AI Prediction](https://cloud.google.com/vertex-ai/docs/predictions/get-predictions).
 

notebooks/official/custom/SDK_Custom_Container_Prediction.ipynb

@@ -58,7 +58,9 @@
       "source": [
         "## Overview\n",
         "\n",
-        "This tutorial walks you through building a custom container to serve a scikit-learn model on Vertex AI. You use the FastAPI Python web server framework to create a prediction and health endpoint. You also incorporate a pre-processor from training pipeline into your online serving application."
+        "This tutorial walks you through building a custom container to serve a scikit-learn model on Vertex AI. You use the FastAPI Python web server framework to create a prediction and health endpoint. You also incorporate a pre-processor from training pipeline into your online serving application.\n",
+        "\n",
+        "Learn more about [Custom training](https://cloud.google.com/vertex-ai/docs/training/custom-training) and [Vertex AI Prediction](https://cloud.google.com/vertex-ai/docs/predictions/get-predictions)."
       ]
     },
     {

notebooks/official/custom/SDK_FBProphet_Forecasting_Online.ipynb

@@ -62,7 +62,9 @@
         "\n",
         "This tutorial walks you through building a custom container to serve a facebook prophet model on Vertex AI. You use the FastAPI Python web server framework to create a prediction endpoint. This notebook is a modified version of an example on [serving a scikit-learn model on Vertex AI](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/sdk/SDK_Custom_Container_Prediction.ipynb).\n",
         "\n",
-        "Learn more about serving an FBProphet model from this [article on testdriven.io: Deploying and Hosting a Machine Learning Model with FastAPI and Heroku](https://testdriven.io/blog/fastapi-machine-learning/).\n"
+        "Learn more about serving an FBProphet model from this [article on testdriven.io: Deploying and Hosting a Machine Learning Model with FastAPI and Heroku](https://testdriven.io/blog/fastapi-machine-learning/).\n",
+        "\n",
+        "Learn more about [Custom training](https://cloud.google.com/vertex-ai/docs/training/custom-training) and [Vertex AI Prediction](https://cloud.google.com/vertex-ai/docs/predictions/get-predictions).\n"
       ]
     },
     {

notebooks/official/custom/custom-tabular-bq-managed-dataset.ipynb

@@ -61,7 +61,9 @@
         "## Overview\n",
         "\n",
         "\n",
-        "This tutorial demonstrates how to use the Vertex AI SDK for Python to train and deploy a custom tabular classification model for online prediction."
+        "This tutorial demonstrates how to use the Vertex AI SDK for Python to train and deploy a custom tabular classification model for online prediction.\n",
+        "\n",
+        "Learn more about [Vertex AI Training](https://cloud.google.com/vertex-ai/docs/training/custom-training)."
       ]
     },
     {
@@ -145,8 +147,7 @@
         "# Install the packages\n",
         "! pip3 install --upgrade google-cloud-aiplatform \\\n",
         "                        google-cloud-storage \\\n",
-        "                        google-cloud-bigquery \\\n",
-        "                        pyarrow"
+        "                        'google-cloud-bigquery[pandas]'"
       ]
     },
     {
@@ -630,12 +631,12 @@
         "    df_train_x, df_train_y = df_train, df_train.pop(LABEL_COLUMN)\n",
         "    df_validation_x, df_validation_y = df_validation, df_validation.pop(LABEL_COLUMN)\n",
         "\n",
-        "    y_train = np.asarray(df_train_y).astype(\"float32\")\n",
-        "    y_validation = np.asarray(df_validation_y).astype(\"float32\")\n",
+        "    y_train = tf.convert_to_tensor(np.asarray(df_train_y).astype(\"float32\"))\n",
+        "    y_validation = tf.convert_to_tensor(np.asarray(df_validation_y).astype(\"float32\"))\n",
         "\n",
         "    # Convert to numpy representation\n",
-        "    x_train = np.asarray(df_train_x) \n",
-        "    x_test = np.asarray(df_validation_x)\n",
+        "    x_train = tf.convert_to_tensor(np.asarray(df_train_x).astype(\"float32\"))\n",
+        "    x_test = tf.convert_to_tensor(np.asarray(df_validation_x).astype(\"float32\"))\n",
         "\n",
         "    # Convert to one-hot representation\n",
         "    num_species = len(df_train_y.unique())\n",
@@ -733,7 +734,7 @@
         "    display_name=JOB_NAME,\n",
         "    script_path=\"task.py\",\n",
         "    container_uri=\"us-docker.pkg.dev/vertex-ai/training/tf-cpu.2-8:latest\",\n",
-        "    requirements=[\"google-cloud-bigquery>=2.20.0\", \"db-dtypes\"],\n",
+        "    requirements=[\"google-cloud-bigquery[pandas]\", \"protobuf<3.20.0\"],\n",
         "    model_serving_container_image_uri=\"us-docker.pkg.dev/vertex-ai/prediction/tf2-cpu.2-8:latest\",\n",
         ")\n",
         "\n",

notebooks/official/custom/custom_training_tensorboard_profiler.ipynb

@@ -61,7 +61,9 @@
       "source": [
         "## Overview\n",
         "\n",
-        "Vertex AI TensorBoard Profiler lets you monitor and optimize your model training performance by helping you understand the resource consumption of training operations. This tutorial demonstrates how to enable Vertex AI TensorBoard Profiler so you can debug model training performance for your custom training jobs.\n"
+        "Vertex AI TensorBoard Profiler lets you monitor and optimize your model training performance by helping you understand the resource consumption of training operations. This tutorial demonstrates how to enable Vertex AI TensorBoard Profiler so you can debug model training performance for your custom training jobs.\n",
+        "\n",
+        "Learn more about [Vertex AI TensorBoard Profiler](https://cloud.google.com/vertex-ai/docs/experiments/tensorboard-profiler)."
       ]
     },
     {

notebooks/official/custom/get_started_vertex_training_xgboost.ipynb

@@ -0,0 +1,994 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "copyright"
+      },
+      "outputs": [],
+      "source": [
+        "# Copyright 2022 Google LLC\n",
+        "#\n",
+        "# Licensed under the Apache License, Version 2.0 (the \"License\");\n",
+        "# you may not use this file except in compliance with the License.\n",
+        "# You may obtain a copy of the License at\n",
+        "#\n",
+        "#     https://www.apache.org/licenses/LICENSE-2.0\n",
+        "#\n",
+        "# Unless required by applicable law or agreed to in writing, software\n",
+        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
+        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
+        "# See the License for the specific language governing permissions and\n",
+        "# limitations under the License."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "title:generic,gcp"
+      },
+      "source": [
+        "# Get started with Vertex AI Training for XGBoost\n",
+        "\n",
+        "<table align=\"left\">\n",
+        "  <td>\n",
+        "    <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/custom/get_started_vertex_training.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/colab-logo-32px.png\" alt=\"Colab logo\"> Run in Colab\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>\n",
+        "    <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/custom/get_started_vertex_training.ipynb\">\n",
+        "      <img src=\"https://cloud.google.com/ml-engine/images/github-logo-32px.png\" alt=\"GitHub logo\">\n",
+        "      View on GitHub\n",
+        "    </a>\n",
+        "  </td>\n",
+        "  <td>\n",
+        "    <a href=\"https://console.cloud.google.com/vertex-ai/workbench/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/main/notebooks/official/custom/get_started_vertex_training.ipynb\">\n",
+        "        <img src=\"https://lh3.googleusercontent.com/UiNooY4LUgW_oTvpsNhPpQzsstV5W8F7rYgxgGBD85cWJoLmrOzhVs_ksK_vgx40SHs7jCqkTkCk=e14-rj-sc0xffffff-h130-w32\" alt=\"Vertex AI logo\">\n",
+        "      Open in Vertex AI Workbench\n",
+        "    </a>\n",
+        "  </td>\n",
+        "</table>\n",
+        "<br/><br/><br/>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "overview:mlops"
+      },
+      "source": [
+        "## Overview\n",
+        "\n",
+        "This tutorial demonstrates how to use Vertex AI Training for XGBoost models.\n",
+        "\n",
+        "Learn more about [Custom training](https://cloud.google.com/vertex-ai/docs/training/custom-training)."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "objective:mlops,stage2,get_started_vertex_training_xgboost"
+      },
+      "source": [
+        "### Objective\n",
+        "\n",
+        "In this tutorial, you learn how to use `Vertex AI Training` for training a XGBoost custom model.\n",
+        "\n",
+        "This tutorial uses the following Google Cloud ML services:\n",
+        "\n",
+        "- `Vertex AI Training`\n",
+        "- `Vertex AI Model` resource\n",
+        "\n",
+        "The steps performed include:\n",
+        "\n",
+        "- Training using a Python package.\n",
+        "- Report accuracy when hyperparameter tuning.\n",
+        "- Save the model artifacts to Cloud Storage using GCSFuse.\n",
+        "- Create a `Vertex AI Model` resource."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "dataset:iris,lcn"
+      },
+      "source": [
+        "### Dataset\n",
+        "\n",
+        "The dataset used for this tutorial is the [Iris dataset](https://www.tensorflow.org/datasets/catalog/iris) from [TensorFlow Datasets](https://www.tensorflow.org/datasets/catalog/overview). This dataset does not require any feature engineering. The version of the dataset in this tutorial is stored in a public Cloud Storage bucket. The trained model predicts the type of Iris flower species from a class of three species: setosa, virginica, or versicolor."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "4fc0ad661ebb"
+      },
+      "source": [
+        "### Costs \n",
+        "\n",
+        "This tutorial uses billable components of Google Cloud:\n",
+        "\n",
+        "* Vertex AI\n",
+        "* Cloud Storage\n",
+        "\n",
+        "\n",
+        "Learn about [Vertex AI\n",
+        "pricing](https://cloud.google.com/vertex-ai/pricing) and [Cloud Storage\n",
+        "pricing](https://cloud.google.com/storage/pricing), and use the [Pricing\n",
+        "Calculator](https://cloud.google.com/products/calculator/)\n",
+        "to generate a cost estimate based on your projected usage."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "install_mlops"
+      },
+      "source": [
+        "## Installations\n",
+        "\n",
+        "Install the following packages to execute this notebook."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "ncRJ_Dfdox9L"
+      },
+      "outputs": [],
+      "source": [
+        "! pip3 install --upgrade google-cloud-aiplatform  -quiet"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "e9255e3b156f"
+      },
+      "source": [
+        "### Colab Only: Uncomment the following cell to restart the kernel"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "0c0b2427998a"
+      },
+      "outputs": [],
+      "source": [
+        "# Automatically restart kernel after installs so that your environment can access the new packages\n",
+        "# import IPython\n",
+        "\n",
+        "# app = IPython.Application.instance()\n",
+        "# app.kernel.do_shutdown(True)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "435b8e413535"
+      },
+      "source": [
+        "### Before you begin\n",
+        "\n",
+        "#### Set your project ID\n",
+        "\n",
+        "**If you don't know your project ID**, try the following:\n",
+        "-  Run `gcloud config list`\n",
+        "-  Run `gcloud projects list`\n",
+        "-  See the support page: [Locate the project ID](https://support.google.com/googleapi/answer/7014113)"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "be175254a715"
+      },
+      "outputs": [],
+      "source": [
+        "PROJECT_ID = \"[your-project-id]\"  # @param {type:\"string\"}\n",
+        "\n",
+        "# set the project id\n",
+        "! gcloud config set project $PROJECT_ID"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "2e6b8b324ce1"
+      },
+      "source": [
+        "#### Region\n",
+        "\n",
+        "You can also change the `REGION` variable used by Vertex AI. \n",
+        "Learn more about [Vertex AI regions](https://cloud.google.com/vertex-ai/docs/general/locations)."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "region"
+      },
+      "outputs": [],
+      "source": [
+        "REGION = \"us-central1\"  # @param {type: \"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "6c43a8673066"
+      },
+      "source": [
+        "### Authenticate your Google Cloud account\n",
+        "\n",
+        "Depending on your Jupyter environment, you may have to manually authenticate. Follow the relevant instructions below.\n",
+        "\n",
+        "**1. Vertex AI Workbench** \n",
+        "- Do nothing as you are already authenticated.\n",
+        "\n",
+        "**2. Local JupyterLab Instance,** uncomment and run."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "fbc9cd30cc4b"
+      },
+      "outputs": [],
+      "source": [
+        "# ! gcloud auth login"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "cd0da2c26879"
+      },
+      "source": [
+        "**3. Colab,** uncomment and run:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "a336a05c6149"
+      },
+      "outputs": [],
+      "source": [
+        "# from google.colab import auth\n",
+        "# auth.authenticate_user()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0461097edfa5"
+      },
+      "source": [
+        "**4. Service Account or other**\n",
+        "- See how to grant Cloud Storage permissions to your service account at [IAM Ch Examples](https://cloud.google.com/storage/docs/gsutil/commands/iam#ch-examples)."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "e5755d1a554f"
+      },
+      "source": [
+        "### Create a Cloud Storage bucket\n",
+        "\n",
+        "Create a storage bucket to store intermediate artifacts such as datasets."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "d2de92accb67"
+      },
+      "outputs": [],
+      "source": [
+        "BUCKET_URI = f\"gs://your-bucket-name-unique-{PROJECT_ID}\"  # @param {type:\"string\"}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "create_bucket"
+      },
+      "source": [
+        "**Only if your bucket doesn't already exist**: Run the following cell to create your Cloud Storage bucket."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "aO4sKJfFox9R"
+      },
+      "outputs": [],
+      "source": [
+        "! gsutil mb -l $REGION $BUCKET_URI"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "validate_bucket"
+      },
+      "source": [
+        "Finally, validate access to your Cloud Storage bucket by examining its contents:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "yWnghzKFox9S"
+      },
+      "outputs": [],
+      "source": [
+        "! gsutil ls -al $BUCKET_URI"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "setup_vars"
+      },
+      "source": [
+        "### Set up variables\n",
+        "\n",
+        "Next, set up some variables used throughout the tutorial.\n",
+        "### Import libraries and define constants"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "import_aip:mbsdk"
+      },
+      "outputs": [],
+      "source": [
+        "import os\n",
+        "\n",
+        "import google.cloud.aiplatform as aiplatform"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "init_aip:mbsdk"
+      },
+      "source": [
+        "### Initialize Vertex AI SDK for Python\n",
+        "\n",
+        "Initialize the Vertex AI SDK for Python for your project and corresponding bucket."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "JZg2sszQox9T"
+      },
+      "outputs": [],
+      "source": [
+        "aiplatform.init(project=PROJECT_ID, staging_bucket=BUCKET_URI)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "accelerators:training,cpu,prediction,cpu,mbsdk"
+      },
+      "source": [
+        "#### Set hardware accelerators\n",
+        "\n",
+        "You can set hardware accelerators for training and prediction.\n",
+        "\n",
+        "Set the variables `TRAIN_GPU/TRAIN_NGPU` and `DEPLOY_GPU/DEPLOY_NGPU` to use a container image supporting a GPU and the number of GPUs allocated to the virtual machine (VM) instance. For example, to use a GPU container image with 4 NVIDIA Tesla K80 GPUs allocated to each VM, you would specify:\n",
+        "\n",
+        "    (aiplatform.gapic..AcceleratorType.NVIDIA_TESLA_K80, 4)\n",
+        "\n",
+        "\n",
+        "Otherwise specify `(None, None)` to use a container image to run on a CPU.\n",
+        "\n",
+        "Learn more about [hardware accelerator support for your region](https://cloud.google.com/vertex-ai/docs/general/locations#accelerators).\n",
+        "\n",
+        "*Note*: TF releases before 2.3 for GPU support will fail to load the custom model in this tutorial. It is a known issue and fixed in TF 2.3. This is caused by static graph ops that are generated in the serving function. If you encounter this issue on your own custom models, use a container image for TF 2.3 with GPU support."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "cQUrG4Mbox9T"
+      },
+      "outputs": [],
+      "source": [
+        "TRAIN_GPU, TRAIN_NGPU = (None, None)\n",
+        "\n",
+        "DEPLOY_GPU, DEPLOY_NGPU = (None, None)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "container:training,prediction,xgboost"
+      },
+      "source": [
+        "#### Set pre-built containers\n",
+        "\n",
+        "Set the pre-built Docker container image for training and prediction.\n",
+        "\n",
+        "\n",
+        "For the latest list, see [Pre-built containers for training](https://cloud.google.com/ai-platform-unified/docs/training/pre-built-containers).\n",
+        "\n",
+        "\n",
+        "For the latest list, see [Pre-built containers for prediction](https://cloud.google.com/ai-platform-unified/docs/predictions/pre-built-containers)."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "XujRA5ueox9U"
+      },
+      "outputs": [],
+      "source": [
+        "TRAIN_VERSION = \"xgboost-cpu.1-1\"\n",
+        "DEPLOY_VERSION = \"xgboost-cpu.1-1\"\n",
+        "\n",
+        "TRAIN_IMAGE = \"{}-docker.pkg.dev/vertex-ai/training/{}:latest\".format(\n",
+        "    REGION.split(\"-\")[0], TRAIN_VERSION\n",
+        ")\n",
+        "DEPLOY_IMAGE = \"{}-docker.pkg.dev/vertex-ai/prediction/{}:latest\".format(\n",
+        "    REGION.split(\"-\")[0], DEPLOY_VERSION\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "machine:training"
+      },
+      "source": [
+        "#### Set machine type\n",
+        "\n",
+        "Next, set the machine type to use for training.\n",
+        "\n",
+        "- Set the variable `TRAIN_COMPUTE` to configure  the compute resources for the VMs you will use for for training.\n",
+        " - `machine type`\n",
+        "     - `n1-standard`: 3.75GB of memory per vCPU.\n",
+        "     - `n1-highmem`: 6.5GB of memory per vCPU\n",
+        "     - `n1-highcpu`: 0.9 GB of memory per vCPU\n",
+        " - `vCPUs`: number of \\[2, 4, 8, 16, 32, 64, 96 \\]\n",
+        "\n",
+        "*Note: The following is not supported for training:*\n",
+        "\n",
+        " - `standard`: 2 vCPUs\n",
+        " - `highcpu`: 2, 4 and 8 vCPUs\n",
+        "\n",
+        "*Note: You may also use n2 and e2 machine types for training and deployment, but they do not support GPUs*."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "UMPFgENkox9U"
+      },
+      "outputs": [],
+      "source": [
+        "TRAIN_COMPUTE = \"n1-standard-4\"\n",
+        "print(\"Train machine type\", TRAIN_COMPUTE)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "xgboost_intro"
+      },
+      "source": [
+        "## Introduction to XGBoost training\n",
+        "\n",
+        "Once you have trained a XGBoost model, you will want to save it at a Cloud Storage location, so it can subsequently be uploaded to a `Vertex AI Model` resource.\n",
+        "The XGBoost package does not have support to save the model to a Cloud Storage location. Instead, you will do the following steps to save to a Cloud Storage location.\n",
+        "\n",
+        "1. Save the in-memory model to the local filesystem (e.g., model.bst).\n",
+        "2. Use gsutil to copy the local copy to the specified Cloud Storage location.\n",
+        "\n",
+        "*Note*: You can do hyperparameter tuning with a XGBoost model."
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "examine_training_package:xgboost"
+      },
+      "source": [
+        "### Examine the training package\n",
+        "\n",
+        "#### Package layout\n",
+        "\n",
+        "Before you start the training, you will look at how a Python package is assembled for a custom training job. When unarchived, the package contains the following directory/file layout.\n",
+        "\n",
+        "- PKG-INFO\n",
+        "- README.md\n",
+        "- setup.cfg\n",
+        "- setup.py\n",
+        "- trainer\n",
+        "  - \\_\\_init\\_\\_.py\n",
+        "  - task.py\n",
+        "\n",
+        "The files `setup.cfg` and `setup.py` are the instructions for installing the package into the operating environment of the Docker image.\n",
+        "\n",
+        "The file `trainer/task.py` is the Python script for executing the custom training job. *Note*, when we referred to it in the worker pool specification, we replace the directory slash with a dot (`trainer.task`) and dropped the file suffix (`.py`).\n",
+        "\n",
+        "#### Package Assembly\n",
+        "\n",
+        "In the following cells, you will assemble the training package."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "f4wS4eISox9V"
+      },
+      "outputs": [],
+      "source": [
+        "# Make folder for Python training script\n",
+        "! rm -rf custom\n",
+        "! mkdir custom\n",
+        "\n",
+        "# Add package information\n",
+        "! touch custom/README.md\n",
+        "\n",
+        "setup_cfg = \"[egg_info]\\n\\ntag_build =\\n\\ntag_date = 0\"\n",
+        "! echo \"$setup_cfg\" > custom/setup.cfg\n",
+        "\n",
+        "setup_py = \"import setuptools\\n\\nsetuptools.setup(\\n\\n    install_requires=[\\n\\n        'cloudml-hypertune',\\n\\n    ],\\n\\n    packages=setuptools.find_packages())\"\n",
+        "! echo \"$setup_py\" > custom/setup.py\n",
+        "\n",
+        "pkg_info = \"Metadata-Version: 1.0\\n\\nName: Iris tabular classification\\n\\nVersion: 0.0.0\\n\\nSummary: Demostration training script\\n\\nHome-page: www.google.com\\n\\nAuthor: Google\\n\\nAuthor-email: aferlitsch@google.com\\n\\nLicense: Public\\n\\nDescription: Demo\\n\\nPlatform: Vertex\"\n",
+        "! echo \"$pkg_info\" > custom/PKG-INFO\n",
+        "\n",
+        "# Make the training subfolder\n",
+        "! mkdir custom/trainer\n",
+        "! touch custom/trainer/__init__.py"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "taskpy_contents:iris,xgboost"
+      },
+      "source": [
+        "### Create the task script for the Python training package\n",
+        "\n",
+        "Next, you create the `task.py` script for driving the training package. Some noteable steps include:\n",
+        "\n",
+        "- Command-line arguments:\n",
+        "    - `model-dir`: The location to save the trained model. When using Vertex AI custom training, the location will be specified in the environment variable: `AIP_MODEL_DIR`,\n",
+        "    - `dataset_data_url`: The location of the training data to download.\n",
+        "    - `dataset_labels_url`: The location of the training labels to download.\n",
+        "    - `boost-rounds`: Tunable hyperparameter\n",
+        "- Data preprocessing (`get_data()`):\n",
+        "    - Download the dataset and split into training and test.\n",
+        "- Training (`train_model()`):\n",
+        "    - Trains the model\n",
+        "- Evaluation (`evaluate_model()`):\n",
+        "    - Evaluates the model.\n",
+        "    - If hyperparameter tuning, reports the metric for accuracy.\n",
+        "- Model artifact saving\n",
+        "    - Saves the model artifacts and evaluation metrics where the Cloud Storage location specified by `model-dir`.\n",
+        "    \n",
+        "*Note:* The training script uses GCSFuse which mounts the Cloud Storage bucket as a network filesystem, allowing the script to perform filesystem operations (e.g., read and write) within a Cloud Storage bucket."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "WiSnFuDoox9W"
+      },
+      "outputs": [],
+      "source": [
+        "%%writefile custom/trainer/task.py\n",
+        "import datetime\n",
+        "import os\n",
+        "import subprocess\n",
+        "import sys\n",
+        "import pandas as pd\n",
+        "import xgboost as xgb\n",
+        "import hypertune\n",
+        "import argparse\n",
+        "import logging\n",
+        "import numpy as np\n",
+        "\n",
+        "from sklearn.model_selection import train_test_split\n",
+        "from sklearn.metrics import accuracy_score\n",
+        "\n",
+        "parser = argparse.ArgumentParser()\n",
+        "parser.add_argument('--model-dir', dest='model_dir',\n",
+        "                    default=os.getenv('AIP_MODEL_DIR'), type=str, help='Model dir.')\n",
+        "parser.add_argument(\"--dataset-data-url\", dest=\"dataset_data_url\",\n",
+        "                    type=str, help=\"Download url for the training data.\")\n",
+        "parser.add_argument(\"--dataset-labels-url\", dest=\"dataset_labels_url\",\n",
+        "                    type=str, help=\"Download url for the training data labels.\")\n",
+        "parser.add_argument(\"--boost-rounds\", dest=\"boost_rounds\",\n",
+        "                    default=20, type=int, help=\"Number of boosted rounds\")\n",
+        "args = parser.parse_args()\n",
+        "\n",
+        "logging.getLogger().setLevel(logging.INFO)\n",
+        "\n",
+        "def get_data():\n",
+        "    logging.info(\"Downloading training data and labelsfrom: {}, {}\".format(args.dataset_data_url, args.dataset_labels_url))\n",
+        "    # gsutil outputs everything to stderr so we need to divert it to stdout.\n",
+        "    subprocess.check_call(['gsutil', 'cp', args.dataset_data_url, 'data.csv'], stderr=sys.stdout)\n",
+        "    # gsutil outputs everything to stderr so we need to divert it to stdout.\n",
+        "    subprocess.check_call(['gsutil', 'cp', args.dataset_labels_url, 'labels.csv'], stderr=sys.stdout)\n",
+        "\n",
+        "\n",
+        "    # Load data into pandas, then use `.values` to get NumPy arrays\n",
+        "    data = pd.read_csv('data.csv').values\n",
+        "    labels = pd.read_csv('labels.csv').values\n",
+        "\n",
+        "    # Convert one-column 2D array into 1D array for use with XGBoost\n",
+        "    labels = labels.reshape((labels.size,))\n",
+        "\n",
+        "    train_data, test_data, train_labels, test_labels = train_test_split(data, labels, test_size=0.2, random_state=7)\n",
+        "\n",
+        "    # Load data into DMatrix object\n",
+        "    dtrain = xgb.DMatrix(train_data, label=train_labels)\n",
+        "    return dtrain, test_data, test_labels\n",
+        "\n",
+        "def train_model(dtrain):\n",
+        "    logging.info(\"Start training ...\")\n",
+        "    # Train XGBoost model\n",
+        "    params = {\n",
+        "        'objective': 'multi:softprob',\n",
+        "        'num_class': 3\n",
+        "    }\n",
+        "    model = xgb.train(params, dtrain, num_boost_round=args.boost_rounds)\n",
+        "    logging.info(\"Training completed\")\n",
+        "    return model\n",
+        "\n",
+        "def evaluate_model(model, test_data, test_labels):\n",
+        "    dtest = xgb.DMatrix(test_data)\n",
+        "    pred = model.predict(dtest)\n",
+        "    predictions = [np.around(value) for value in pred]\n",
+        "    # evaluate predictions\n",
+        "    try:\n",
+        "        accuracy = accuracy_score(test_labels, predictions)\n",
+        "    except:\n",
+        "        accuracy = 0.0\n",
+        "    logging.info(f\"Evaluation completed with model accuracy: {accuracy}\")\n",
+        "\n",
+        "    # report metric for hyperparameter tuning\n",
+        "    hpt = hypertune.HyperTune()\n",
+        "    hpt.report_hyperparameter_tuning_metric(\n",
+        "        hyperparameter_metric_tag='accuracy',\n",
+        "        metric_value=accuracy\n",
+        "    )\n",
+        "    return accuracy\n",
+        "\n",
+        "\n",
+        "dtrain, test_data, test_labels = get_data()\n",
+        "model = train_model(dtrain)\n",
+        "accuracy = evaluate_model(model, test_data, test_labels)\n",
+        "\n",
+        "# GCSFuse conversion\n",
+        "gs_prefix = 'gs://'\n",
+        "gcsfuse_prefix = '/gcs/'\n",
+        "if args.model_dir.startswith(gs_prefix):\n",
+        "    args.model_dir = args.model_dir.replace(gs_prefix, gcsfuse_prefix)\n",
+        "    dirpath = os.path.split(args.model_dir)[0]\n",
+        "    if not os.path.isdir(dirpath):\n",
+        "        os.makedirs(dirpath)\n",
+        "\n",
+        "# Export the classifier to a file\n",
+        "gcs_model_path = os.path.join(args.model_dir, 'model.bst')\n",
+        "logging.info(\"Saving model artifacts to {}\". format(gcs_model_path))\n",
+        "model.save_model(gcs_model_path)\n",
+        "\n",
+        "logging.info(\"Saving metrics to {}/metrics.json\". format(args.model_dir))\n",
+        "gcs_metrics_path = os.path.join(args.model_dir, 'metrics.json')\n",
+        "with open(gcs_metrics_path, \"w\") as f:\n",
+        "    f.write(f\"{'accuracy: {accuracy}'}\")"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "tarball_training_script"
+      },
+      "source": [
+        "#### Store training script on your Cloud Storage bucket\n",
+        "\n",
+        "Next, you package the training folder into a compressed tar ball, and then store it in your Cloud Storage bucket."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "dnmdycf6ox9X"
+      },
+      "outputs": [],
+      "source": [
+        "! rm -f custom.tar custom.tar.gz\n",
+        "! tar cvf custom.tar custom\n",
+        "! gzip custom.tar\n",
+        "! gsutil cp custom.tar.gz $BUCKET_URI/trainer_iris.tar.gz"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "create_custom_pp_training_job:mbsdk"
+      },
+      "source": [
+        "### Create and run custom training job\n",
+        "\n",
+        "\n",
+        "To train a custom model, you perform two steps: 1) create a custom training job, and 2) run the job.\n",
+        "\n",
+        "#### Create custom training job\n",
+        "\n",
+        "A custom training job is created with the `CustomTrainingJob` class, with the following parameters:\n",
+        "\n",
+        "- `display_name`: The human readable name for the custom training job.\n",
+        "- `container_uri`: The training container image.\n",
+        "\n",
+        "- `python_package_gcs_uri`: The location of the Python training package as a tarball.\n",
+        "- `python_module_name`: The relative path to the training script in the Python package.\n",
+        "- `model_serving_container_uri`: The container image for deploying the model.\n",
+        "\n",
+        "*Note:* There is no requirements parameter. You specify any requirements in the `setup.py` script in your Python package."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "rVEMz1xqox9X"
+      },
+      "outputs": [],
+      "source": [
+        "DISPLAY_NAME = \"iris\"\n",
+        "\n",
+        "job = aiplatform.CustomPythonPackageTrainingJob(\n",
+        "    display_name=DISPLAY_NAME,\n",
+        "    python_package_gcs_uri=f\"{BUCKET_URI}/trainer_iris.tar.gz\",\n",
+        "    python_module_name=\"trainer.task\",\n",
+        "    container_uri=TRAIN_IMAGE,\n",
+        "    model_serving_container_image_uri=DEPLOY_IMAGE,\n",
+        "    project=PROJECT_ID,\n",
+        ")"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "prepare_custom_cmdargs:iris,xgboost"
+      },
+      "source": [
+        "### Prepare your command-line arguments\n",
+        "\n",
+        "Now define the command-line arguments for your custom training container:\n",
+        "\n",
+        "- `args`: The command-line arguments to pass to the executable that is set as the entry point into the container.\n",
+        "  - `--model-dir` : For our demonstrations, we use this command-line argument to specify where to store the model artifacts.\n",
+        "      - direct: You pass the Cloud Storage location as a command line argument to your training script (set variable `DIRECT = True`), or\n",
+        "      - indirect: The service passes the Cloud Storage location as the environment variable `AIP_MODEL_DIR` to your training script (set variable `DIRECT = False`). In this case, you tell the service the model artifact location in the job specification.\n",
+        "  - `--dataset-data-url`: The location of the training data to download.\n",
+        "  - `--dataset-labels-url`: The location of the training labels to download.\n",
+        "  - `--boost-rounds`: Tunable hyperparameter."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "AoUfpBqVox9Y"
+      },
+      "outputs": [],
+      "source": [
+        "MODEL_DIR = \"{}/{}\".format(BUCKET_URI, \"model\")\n",
+        "DATASET_DIR = \"gs://cloud-samples-data/ai-platform/iris\"\n",
+        "\n",
+        "ROUNDS = 20\n",
+        "\n",
+        "DIRECT = False\n",
+        "if DIRECT:\n",
+        "    CMDARGS = [\n",
+        "        \"--dataset-data-url=\" + DATASET_DIR + \"/iris_data.csv\",\n",
+        "        \"--dataset-labels-url=\" + DATASET_DIR + \"/iris_target.csv\",\n",
+        "        \"--boost-rounds=\" + str(ROUNDS),\n",
+        "        \"--model_dir=\" + MODEL_DIR,\n",
+        "    ]\n",
+        "else:\n",
+        "    CMDARGS = [\n",
+        "        \"--dataset-data-url=\" + DATASET_DIR + \"/iris_data.csv\",\n",
+        "        \"--dataset-labels-url=\" + DATASET_DIR + \"/iris_target.csv\",\n",
+        "        \"--boost-rounds=\" + str(ROUNDS),\n",
+        "    ]"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "run_custom_job:mbsdk"
+      },
+      "source": [
+        "#### Run the custom training job\n",
+        "\n",
+        "Next, you run the custom job to start the training job by invoking the method `run`, with the following parameters:\n",
+        "\n",
+        "- `model_display_name`: The human readable name for the `Model` resource.\n",
+        "- `args`: The command-line arguments to pass to the training script.\n",
+        "- `replica_count`: The number of compute instances for training (replica_count = 1 is single node training).\n",
+        "- `machine_type`: The machine type for the compute instances.\n",
+        "- `accelerator_type`: The hardware accelerator type.\n",
+        "- `accelerator_count`: The number of accelerators to attach to a worker replica.\n",
+        "- `base_output_dir`: The Cloud Storage location to write the model artifacts to.\n",
+        "- `sync`: Whether to block until completion of the job."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "JCruQq1aox9Y"
+      },
+      "outputs": [],
+      "source": [
+        "if TRAIN_GPU:\n",
+        "    model = job.run(\n",
+        "        model_display_name=\"iris\",\n",
+        "        args=CMDARGS,\n",
+        "        replica_count=1,\n",
+        "        machine_type=TRAIN_COMPUTE,\n",
+        "        accelerator_type=TRAIN_GPU.name,\n",
+        "        accelerator_count=TRAIN_NGPU,\n",
+        "        base_output_dir=MODEL_DIR,\n",
+        "        sync=False,\n",
+        "    )\n",
+        "else:\n",
+        "    model = job.run(\n",
+        "        model_display_name=\"iris\",\n",
+        "        args=CMDARGS,\n",
+        "        replica_count=1,\n",
+        "        machine_type=TRAIN_COMPUTE,\n",
+        "        base_output_dir=MODEL_DIR,\n",
+        "        sync=False,\n",
+        "    )\n",
+        "\n",
+        "model_path_to_deploy = MODEL_DIR"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "list_job"
+      },
+      "source": [
+        "### List a custom training job"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "KBM_KLMSox9Y"
+      },
+      "outputs": [],
+      "source": [
+        "_job = job.list(filter=f\"display_name={DISPLAY_NAME}\")\n",
+        "print(_job)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "custom_job_wait:mbsdk"
+      },
+      "source": [
+        "### Wait for completion of custom training job\n",
+        "\n",
+        "Next, wait for the custom training job to complete. Alternatively, one can set the parameter `sync` to `True` in the `run()` method to block until the custom training job is completed."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "lHPMHbSyox9Z"
+      },
+      "outputs": [],
+      "source": [
+        "model.wait()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "delete_job"
+      },
+      "source": [
+        "### Delete a custom training job\n",
+        "\n",
+        "After a training job is completed, you can delete the training job with the method `delete()`.  Prior to completion, a training job can be canceled with the method `cancel()`."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "tlYg7Sp-ox9Z"
+      },
+      "outputs": [],
+      "source": [
+        "job.delete()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "cleanup:mbsdk"
+      },
+      "source": [
+        "# Cleaning up\n",
+        "\n",
+        "To clean up all Google Cloud resources used in this project, you can [delete the Google Cloud\n",
+        "project](https://cloud.google.com/resource-manager/docs/creating-managing-projects#shutting_down_projects) you used for the tutorial.\n",
+        "\n",
+        "Otherwise, you can delete the individual resources you created in this tutorial:\n",
+        "\n",
+        "- Custom Job (Custome Training job is remove in previous step)\n",
+        "- Cloud Storage Bucket"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "JyWy23gDox9a"
+      },
+      "outputs": [],
+      "source": [
+        "delete_bucket = True\n",
+        "\n",
+        "model.delete()\n",
+        "\n",
+        "if delete_bucket or os.getenv(\"IS_TESTING\"):\n",
+        "    ! gsutil rm -r $BUCKET_URI"
+      ]
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "name": "get_started_vertex_training_xgboost.ipynb",
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

notebooks/official/custom/sdk-custom-image-classification-batch.ipynb

@@ -62,7 +62,9 @@
         "## Overview\n",
         "\n",
         "\n",
-        "This tutorial demonstrates how to use the Vertex AI SDK for Python to train and deploy a custom image classification model for batch prediction."
+        "This tutorial demonstrates how to use the Vertex AI SDK for Python to train and deploy a custom image classification model for batch prediction.\n",
+        "\n",
+        "Learn more about [Custom training](https://cloud.google.com/vertex-ai/docs/training/custom-training) and [Vertex AI Batch Prediction](https://cloud.google.com/vertex-ai/docs/tabular-data/classification-regression/get-batch-predictions)."
       ]
     },
     {

notebooks/official/custom/sdk-custom-image-classification-online.ipynb

@@ -62,7 +62,9 @@
         "## Overview\n",
         "\n",
         "\n",
-        "This tutorial demonstrates how to use the Vertex AI SDK for Python to train and deploy a custom image classification model for online prediction."
+        "This tutorial demonstrates how to use the Vertex AI SDK for Python to train and deploy a custom image classification model for online prediction.\n",
+        "\n",
+        "Learn more about [Custom training](https://cloud.google.com/vertex-ai/docs/training/custom-training) and [Vertex AI Prediction](https://cloud.google.com/vertex-ai/docs/predictions/get-predictions)."
       ]
     },
     {

notebooks/official/datasets/README.md

@@ -0,0 +1,38 @@
+
+[Get started with BigQuery datasets](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/datasets/get_started_bq_datasets.ipynb)
+
+```
+Learn how to use `BigQuery` as a dataset for training with `Vertex AI`.
+
+The steps performed include:
+
+- Create a Vertex AI `Dataset` resource from `BigQuery` table -- compatible for `AutoML` training.
+- Extract a copy of the dataset from `BigQuery` to a CSV file in Cloud Storage -- compatible for `AutoML` or custom training.
+- Select rows from a `BigQuery` dataset into a `pandas` dataframe -- compatible for custom training.
+- Select rows from a `BigQuery` dataset into a `tf.data.Dataset` -- compatible for custom training `TensorFlow` models.
+- Select rows from extracted CSV files into a `tf.data.Dataset` -- compatible for custom training `TensorFlow` models.
+- Create a `BigQuery` dataset from CSV files.
+- Extract data from `BigQuery` table into a `DMatrix` -- compatible for custom training `XGBoost` models.
+
+```
+
+   Learn more about [BigQuery Datasets](https://cloud.google.com/bigquery/docs/datasets-intro).
+
+
+[Get started with Vertex AI Data Labeling](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/datasets/get_started_with_data_labeling.ipynb)
+
+```
+Learn how to use the `Vertex AI Data Labeling` service.
+
+The steps performed include:
+
+- Create a Specialist Pool for data labelers.
+- Create a data labeling job.
+- Submit the data labeling job.
+- List data labeling jobs.
+- Cancel a data labeling job.
+
+```
+
+   Learn more about [Vertex AI Data Labeling](https://cloud.google.com/vertex-ai/docs/datasets/data-labeling-job).
+