Compare commits

..
1 Commits
Author SHA1 Message Date
Andrew Ferlitsch 5c49bc6fe7 fix: rename UJ10 2023-04-19 23:51:55 +00:00
62 changed files with 558 additions and 7357 deletions
+13 -52
View File
@@ -17,7 +17,6 @@
import argparse
import pathlib
import os
import execute_changed_notebooks_helper
@@ -40,19 +39,6 @@ parser.add_argument(
help="The path to the file that has newline-limited folders of notebooks that should be tested.",
required=True,
)
parser.add_argument(
"--test_percent",
type=int,
help="The percent of notebooks to be tested (between 1 and 100).",
required=False,
default=100,
)
parser.add_argument(
"--build_id",
type=str,
help="The build id (which may be a Cloud Build job specific or user explicit.",
required=True
)
parser.add_argument(
"--base_branch",
help="The base git branch to diff against to find changed files.",
@@ -121,49 +107,24 @@ parser.add_argument(
default=True,
help="Should run notebooks in parallel.",
)
parser.add_argument(
"--dry_run",
type=str2bool,
default=False,
help="Dry run for testing - no execution",
)
args = parser.parse_args()
changed_notebooks = execute_changed_notebooks_helper.get_changed_notebooks(
notebooks = execute_changed_notebooks_helper.get_changed_notebooks(
test_paths_file=args.test_paths_file,
base_branch=args.base_branch,
)
results_bucket = f"{args.artifacts_bucket}"
results_file = f"{args.build_id}.json"
if args.test_percent == 100:
notebooks = changed_notebooks
accumulative_results = {}
else:
accumulative_results = execute_changed_notebooks_helper.load_results(results_bucket, results_file)
notebooks = [changed_notebook for changed_notebook in changed_notebooks if execute_changed_notebooks_helper.select_notebook(changed_notebook, accumulative_results, args.test_percent)]
if args.dry_run:
print("Dry run ...\n")
for notebook in notebooks:
print(f"Would execute: {notebook}")
else:
execute_changed_notebooks_helper.process_and_execute_notebooks(
notebooks=notebooks,
container_uri=args.container_uri,
staging_bucket=args.staging_bucket,
artifacts_bucket=args.artifacts_bucket,
results_file=results_file,
accumulative_results=accumulative_results,
should_parallelize=args.should_parallelize,
timeout=args.timeout,
variable_project_id=args.variable_project_id,
variable_region=args.variable_region,
variable_service_account=args.variable_service_account,
variable_vpc_network=args.variable_vpc_network,
private_pool_id=args.private_pool_id
execute_changed_notebooks_helper.process_and_execute_notebooks(
notebooks=notebooks,
container_uri=args.container_uri,
staging_bucket=args.staging_bucket,
artifacts_bucket=args.artifacts_bucket,
should_parallelize=args.should_parallelize,
timeout=args.timeout,
variable_project_id=args.variable_project_id,
variable_region=args.variable_region,
variable_service_account=args.variable_service_account,
variable_vpc_network=args.variable_vpc_network,
private_pool_id=args.private_pool_id,
)
@@ -21,15 +21,11 @@ import json
import git
import operator
import os
import io
import json
import pathlib
import re
import subprocess
import random
from google.cloud import storage
import utils
from typing import List, Optional, Dict, Any
from typing import List, Optional
from utils import util
import execute_notebook_helper
@@ -69,9 +65,7 @@ def format_timedelta(delta: datetime.timedelta) -> str:
@dataclasses.dataclass
class NotebookExecutionResult:
name: str
path: str
duration: datetime.timedelta
start_time: datetime.datetime
is_pass: bool
log_url: str
output_uri: str
@@ -86,40 +80,6 @@ class NotebookExecutionResult:
else:
return None
def load_results(results_bucket: str,
results_file: str) -> Dict[str,Any]:
'''
Load accumulated notebook test results
'''
print("Loading existing accumulative results ...")
accumulative_results = {}
try:
content = util.download_blob_into_memory(results_bucket, results_file, download_as_text=True)
accumulative_results = json.loads(content)
print(accumulative_results)
except Exception as e:
print(e)
# If there are no accumulative results, an empty dict is returned
return accumulative_results
def select_notebook(changed_notebook: str,
accumulative_results: Dict[str, Any],
test_percent: int) -> bool:
'''
Algorithm to randomly select a notebook, but weight the propbability of selected based on past failures
'''
if changed_notebook in accumulative_results:
pass_count = accumulative_results[changed_notebook]['passed']
fail_count = accumulative_results[changed_notebook]['failed']
else:
pass_count = 1
fail_count = 0
return (random.randint(1, 100) * (1 + (fail_count / (pass_count + fail_count))) < test_percent)
def _process_notebook(
notebook_path: str,
@@ -231,9 +191,7 @@ def process_and_execute_notebook(
result = NotebookExecutionResult(
name=tag,
path=notebook,
duration=datetime.timedelta(seconds=0),
start_time=datetime.datetime.now(),
is_pass=False,
output_uri=notebook_output_uri,
log_url="",
@@ -243,6 +201,7 @@ def process_and_execute_notebook(
)
# TODO: Handle cases where multiple notebooks have the same name
time_start = datetime.datetime.now()
operation = None
try:
# Get the python version for running the notebook if specified
@@ -288,10 +247,9 @@ def process_and_execute_notebook(
# Block and wait for the result
operation_result = operation.result(timeout=timeout_in_seconds)
result.duration = datetime.datetime.now() - result.start_time
result.duration = datetime.datetime.now() - time_start
result.is_pass = True
print(f"{notebook} PASSED in {format_timedelta(result.duration)}.")
except Exception as error:
result.error_message = str(error)
@@ -310,7 +268,7 @@ def process_and_execute_notebook(
except Exception as error:
result.error_message = str(error)
result.duration = datetime.datetime.now() - result.start_time
result.duration = datetime.datetime.now() - time_start
result.is_pass = False
print(
@@ -378,54 +336,12 @@ def get_changed_notebooks(
return notebooks
def _save_results(results: List[NotebookExecutionResult],
accumulative_results: Dict[str,Any],
artifacts_bucket: str,
results_file: str):
artifacts_bucket = artifacts_bucket.replace("gs://", "").split('/')[0]
print("Updating accumulative results ...")
for result in results:
if result.path in accumulative_results:
accumulative_results[result.path]['duration'] = result.duration.total_seconds()
accumulative_results[result.path]['start_time'] = str(result.start_time)
if result.is_pass:
accumulative_results[result.path]['passed'] += 1
else:
accumulative_results[result.path]['failed'] += 1
print(f"updating {result.path}")
else:
if result.is_pass:
pass_count = 1
fail_count = 0
else:
pass_count = 0
fail_count = 1
accumulative_results[result.path] = {
'duration': result.duration.total_seconds(),
'start_time': str(result.start_time),
'passed': pass_count,
'failed': fail_count
}
print(f"adding {result.path}")
print("Saving accumulative results ...")
content = json.dumps(accumulative_results)
client = storage.Client()
bucket = client.get_bucket(artifacts_bucket)
bucket.blob(str(results_file)).upload_from_string(content, 'text/json')
def process_and_execute_notebooks(
notebooks: List[str],
container_uri: str,
staging_bucket: str,
artifacts_bucket: str,
results_file: str,
accumulative_results: List[NotebookExecutionResult],
should_parallelize: bool,
timeout: int,
variable_project_id: str,
@@ -453,10 +369,6 @@ def process_and_execute_notebooks(
Required. The GCS staging bucket to write source code to.
artifacts_bucket (str):
Required. The GCS staging bucket to write executed notebooks to.
results_file (str):
Required: The path to the artifacts bucket to save results
accumulative_results (List):
Required: The in-memory previous accumulative notebook CI/CD test results.
variable_project_id (str):
Required. The value for PROJECT_ID to inject into notebooks.
variable_region (str):
@@ -559,7 +471,7 @@ def process_and_execute_notebooks(
print("=" * 100)
build_id = results_sorted[0].build_id
logs_bucket_name = (results_sorted[0].logs_bucket).replace("gs://", "")
logs_bucket_name = (results_sorted[0].logs_bucket).removeprefix("gs://")
log_file_name = f"log-{build_id}.txt"
log_contents = util.download_blob_into_memory(
@@ -577,11 +489,6 @@ def process_and_execute_notebooks(
else:
print(log_contents)
_save_results(results_sorted,
accumulative_results,
artifacts_bucket,
results_file)
print("\n=== END RESULTS===\n")
total_notebook_duration = functools.reduce(
@@ -36,7 +36,7 @@ steps:
- -c
- |
. workspace/env/bin/activate &&
python3 .cloud-build/execute_changed_notebooks_cli.py --test_paths_file "${_TEST_PATHS_FILE}" --base_branch "${_FORCED_BASE_BRANCH}" --container_uri ${_PYTHON_IMAGE} --staging_bucket ${_GCS_STAGING_BUCKET} --artifacts_bucket ${_GCS_STAGING_BUCKET}/executed_notebooks/PR_${_PR_NUMBER}/BUILD_${BUILD_ID} --variable_project_id ${PROJECT_ID} --variable_region ${_GCP_REGION} --variable_service_account ${_GCP_SERVICE_ACCOUNT} --variable_vpc_network "${_GPC_VPC_NETWORK_NAME}" `if [ ! -z "${_PRIVATE_POOL_NAME}" ]; then echo "--private_pool_id ${_PRIVATE_POOL_NAME}"; fi` --build_id ${BUILD_ID}
python3 .cloud-build/execute_changed_notebooks_cli.py --test_paths_file "${_TEST_PATHS_FILE}" --base_branch "${_FORCED_BASE_BRANCH}" --container_uri ${_PYTHON_IMAGE} --staging_bucket ${_GCS_STAGING_BUCKET} --artifacts_bucket ${_GCS_STAGING_BUCKET}/executed_notebooks/PR_${_PR_NUMBER}/BUILD_${BUILD_ID} --variable_project_id ${PROJECT_ID} --variable_region ${_GCP_REGION} --variable_service_account ${_GCP_SERVICE_ACCOUNT} --variable_vpc_network "${_GPC_VPC_NETWORK_NAME}" `if [ ! -z "${_PRIVATE_POOL_NAME}" ]; then echo "--private_pool_id ${_PRIVATE_POOL_NAME}"; fi`
env:
- 'IS_TESTING=1'
timeout: 86400s
+2 -5
View File
@@ -3,15 +3,12 @@ numpy
jupyter
nbconvert
papermill
pandas
matplotlib
tabulate
google-cloud-aiplatform
google-cloud-storage
google-cloud-build
google-cloud-storage
ratemate
GitPython
tqdm
fsspec
pandas
tqdm
-40
View File
@@ -1,40 +0,0 @@
notebooks/official/training/pytorch_gcs_data_training.ipynb
notebooks/official/custom/custom_training_tensorboard_profiler.ipynb
notebooks/official/custom/custom-tabular-bq-managed-dataset.ipynb
notebooks/official/ml_metadata/sdk-metric-parameter-tracking-for-locally-trained-models.ipynb
notebooks/official/ml_metadata/sdk-metric-parameter-tracking-for-custom-jobs.ipynb
notebooks/official/tabnet/tabnet_vertex_tutorial.ipynb
notebooks/official/tabnet/get_started_with_tabnet.ipynb
notebooks/official/pipelines/google_cloud_pipeline_components_automl_text.ipynb
notebooks/official/pipelines/multicontender_vs_champion_deployment_method.ipynb
notebooks/official/pipelines/google_cloud_pipeline_components_automl_images.ipynb
notebooks/official/pipelines/rapid_prototyping_bqml_automl.ipynb
notebooks/official/pipelines/challenger_vs_blessed_deployment_method.ipynb
notebooks/official/matching_engine/sdk_matching_engine_create_stack_overflow_embeddings.ipynb
notebooks/official/matching_engine/sdk_matching_engine_for_indexing.ipynb
notebooks/official/matching_engine/sdk_matching_engine_create_text_to_image_embeddings.ipynb
notebooks/official/explainable_ai/sdk_custom_image_classification_online_explain.ipynb
notebooks/official/explainable_ai/xai_image_classification_feature_attributions.ipynb
notebooks/official/explainable_ai/sdk_custom_image_classification_batch_explain.ipynb
notebooks/official/tabular_workflows/tabnet_on_vertex_pipelines.ipynb
notebooks/official/model_registry/get_started_with_model_registry.ipynb
notebooks/official/model_registry/bqml_vertexai_model_registry.ipynb
notebooks/official/sdk/SDK_Custom_Training_Python_Package_Managed_Text_Dataset_Tensorflow_Serving_Container.ipynb
notebooks/official/model_monitoring/batch_prediction_model_monitoring.ipynb
notebooks/official/model_monitoring/get_started_with_model_monitoring_setup.ipynb
notebooks/official/model_monitoring/get_started_with_model_monitoring_custom.ipynb
notebooks/official/model_monitoring/get_started_with_model_monitoring_custom_tf_serving.ipynb
notebooks/official/model_monitoring/model_monitoring.ipynb
notebooks/official/tensorboard/tensorboard_profiler_custom_training_with_prebuilt_container.ipynb
notebooks/official/tensorboard/tensorboard_hyperparameter_tuning_with_hparams.ipynb
notebooks/official/tensorboard/tensorboard_profiler_custom_training.ipynb
notebooks/official/model_evaluation/custom_tabular_regression_model_evaluation.ipynb
notebooks/official/model_evaluation/custom_tabular_classification_model_evaluation.ipynb
notebooks/official/model_evaluation/automl_video_classification_model_evaluation.ipynb
notebooks/official/experiments/comparing_local_trained_models.ipynb
notebooks/official/automl/automl_image_classification_online_online_prediction.ipynb
notebooks/official/automl/automl-text-classification.ipynb
notebooks/official/automl/sdk_automl_video_object_tracking_batch.ipynb
notebooks/official/feature_store/sdk-feature-store-pandas.ipynb
notebooks/official/prediction/custom_batch_prediction_feature_filter.ipynb
notebooks/official/prediction/pytorch_image_classification_with_prebuilt_serving_containers.ipynb
-80
View File
@@ -1,80 +0,0 @@
notebooks/official/training/hyperparameter_tuning_tensorflow.ipynb
notebooks/official/training/get_started_with_vertex_distributed_training.ipynb
notebooks/official/training/hyperparameter_tuning_xgboost.ipynb
notebooks/official/training/multi_node_ddp_gloo_vertex_training_with_custom_container.ipynb
notebooks/official/training/distributed_hyperparameter_tuning.ipynb
notebooks/official/training/pytorch-text-sentiment-classification-custom-train-deploy.ipynb
notebooks/official/training/xgboost_data_parallel_training_on_cpu_using_dask.ipynb
notebooks/official/training/multi_node_ddp_nccl_vertex_training_with_custom_container.ipynb
notebooks/official/bigquery_ml/get_started_with_bqml_training.ipynb
notebooks/official/bigquery_ml/bqml-online-prediction.ipynb
notebooks/official/custom/custom_training_container_and_model_registry.ipynb
notebooks/official/custom/sdk-custom-image-classification-online.ipynb
notebooks/official/custom/sdk-custom-image-classification-batch.ipynb
notebooks/official/custom/SDK_FBProphet_Forecasting_Online.ipynb
notebooks/official/custom/get_started_vertex_training_xgboost.ipynb
notebooks/official/custom/get_started_with_vertex_endpoint_and_shared_vm.ipynb
notebooks/official/custom/SDK_Custom_Container_Prediction.ipynb
notebooks/official/reduction_server/pytorch_distributed_training_reduction_server.ipynb
notebooks/official/tabnet/ai-explanations-tabnet-algorithm.ipynb
notebooks/official/vizier/get_started_vertex_vizier.ipynb
notebooks/official/vizier/gapic-vizier-multi-objective-optimization.ipynb
notebooks/official/pipelines/get_started_with_hpt_pipeline_components.ipynb
notebooks/official/pipelines/google_cloud_pipeline_components_automl_tabular.ipynb
notebooks/official/pipelines/custom_tabular_train_batch_pred_bq_pipeline.ipynb
notebooks/official/pipelines/metrics_viz_run_compare_kfp.ipynb
notebooks/official/pipelines/google_cloud_pipeline_components_model_upload_predict_evaluate.ipynb
notebooks/official/pipelines/google_cloud_pipeline_components_model_train_upload_deploy.ipynb
notebooks/official/pipelines/get_started_with_machine_management.ipynb
notebooks/official/pipelines/custom_model_training_and_batch_prediction.ipynb
notebooks/official/pipelines/control_flow_kfp.ipynb
notebooks/official/pipelines/lightweight_functions_component_io_kfp.ipynb
notebooks/official/pipelines/google_cloud_pipeline_components_bqml_text.ipynb
notebooks/official/pipelines/pipelines_intro_kfp.ipynb
notebooks/official/pipelines/automl_tabular_classification_beans.ipynb
notebooks/official/pipelines/google_cloud_pipeline_components_dataproc_tabular.ipynb
notebooks/official/explainable_ai/sdk_automl_tabular_classification_online_explain.ipynb
notebooks/official/explainable_ai/sdk_custom_tabular_regression_online_explain.ipynb
notebooks/official/explainable_ai/sdk_automl_tabular_binary_classification_batch_explain.ipynb
notebooks/official/explainable_ai/sdk_custom_tabular_regression_online_explain_get_metadata.ipynb
notebooks/official/explainable_ai/sdk_custom_tabular_regression_batch_explain.ipynb
notebooks/official/tabular_workflows/prophet_on_vertex_pipelines.ipynb
notebooks/official/tabular_workflows/wide_and_deep_on_vertex_pipelines.ipynb
notebooks/official/sdk/SDK_AutoML_Video_Classification.ipynb
notebooks/official/model_monitoring/get_started_with_model_monitoring_automl.ipynb
notebooks/official/model_monitoring/get_started_with_model_monitoring_automl_image_batch.ipynb
notebooks/official/model_monitoring/get_started_with_model_monitoring_automl_image_online.ipynb
notebooks/official/model_monitoring/get_started_with_model_monitoring_xgboost.ipynb
notebooks/official/tensorboard/tensorboard_custom_training_with_custom_container.ipynb
notebooks/official/tensorboard/tensorboard_custom_training_with_prebuilt_container.ipynb
notebooks/official/tensorboard/tensorboard_vertex_ai_pipelines_integration.ipynb
notebooks/official/model_evaluation/automl_text_classification_model_evaluation.ipynb
notebooks/official/model_evaluation/get_started_with_custom_model_evaluation_import.ipynb
notebooks/official/model_evaluation/automl_tabular_classification_model_evaluation.ipynb
notebooks/official/model_evaluation/automl_tabular_regression_model_evaluation.ipynb
notebooks/official/experiments/get_started_with_vertex_experiments.ipynb
notebooks/official/experiments/comparing_pipeline_runs.ipynb
notebooks/official/experiments/get_started_with_vertex_experiments_autologging.ipynb
notebooks/official/experiments/build_model_experimentation_lineage_with_prebuild_code.ipynb
notebooks/official/experiments/delete_outdated_tensorboard_experiments.ipynb
notebooks/official/automl/sdk_automl_tabular_regression_batch_bq.ipynb
notebooks/official/automl/sdk_automl_text_sentiment_analysis_online.ipynb
notebooks/official/automl/sdk_automl_text_entity_extraction_online.ipynb
notebooks/official/automl/sdk_automl_forecasting_hierarchical_batch.ipynb
notebooks/official/automl/automl_text_entity_extraction_batch_prediction.ipynb
notebooks/official/automl/automl_image_classification_batch_prediction.ipynb
notebooks/official/automl/automl_text_sentiment_analysis_batch_prediction.ipynb
notebooks/official/automl/sdk_automl_tabular_regression_online_bq.ipynb
notebooks/official/automl/get_started_automl_training.ipynb
notebooks/official/automl/automl-tabular-classification.ipynb
notebooks/official/automl/automl_image_object_detection_export_edge.ipynb
notebooks/official/automl/sdk_automl_image_object_detection_batch.ipynb
notebooks/official/automl/automl_tabular_on_vertex_pipelines.ipynb
notebooks/official/automl/sdk_automl_video_classification_batch.ipynb
notebooks/official/automl/sdk_automl_video_action_recognition_batch.ipynb
notebooks/official/automl/sdk_automl_tabular_forecasting_batch.ipynb
notebooks/official/automl/automl_image_object_detection_online_prediction.ipynb
notebooks/official/automl/automl_forecasting_bqml_arima_plus_comparison.ipynb
notebooks/official/datasets/get_started_bq_datasets.ipynb
notebooks/official/datasets/get_started_with_data_labeling.ipynb
notebooks/official/feature_store/feature_store_streaming_ingestion_sdk.ipynb
-46
View File
@@ -1,46 +0,0 @@
# grep PASSED tests.txt | cut -c 10-100 >passed.txt
import os
repo_dir = '/home/jupyter/vertex-ai-samples/'
repo_dir_len = len(repo_dir)
official_dir = repo_dir + 'notebooks/official'
entries = os.scandir(official_dir)
folders = []
for entry in entries:
if entry.is_dir():
folders.append(entry.path)
# Passing
with open('passed.txt', 'r') as pass_file:
notebook_names = pass_file.readlines()
notebooks = []
for folder in folders:
entries = os.scandir(folder)
for entry in entries:
for notebook in notebook_names:
if entry.name == notebook.rstrip():
notebooks.append(entry.path[repo_dir_len:])
with open('passing_tests.txt', 'w') as f:
for notebook in notebooks:
f.write(notebook + '\n')
# Failing
with open('failed.txt', 'r') as fail_file:
notebook_names = fail_file.readlines()
notebooks = []
for folder in folders:
entries = os.scandir(folder)
for entry in entries:
for notebook in notebook_names:
if entry.name == notebook.rstrip():
notebooks.append(entry.path[repo_dir_len:])
with open('failing_tests.txt', 'w') as f:
for notebook in notebooks:
f.write(notebook + '\n')
-1
View File
@@ -56,4 +56,3 @@
/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/model_garden/model_garden_pytorch_blip2.ipynb @xiangxu-google
/notebooks/community/model_garden/model_garden_pytorch_detectron2.ipynb @lavraicse
@@ -196,7 +196,7 @@
" USER_FLAG = \"--user\"\n",
"\n",
"! pip3 install {USER_FLAG} --upgrade --quiet google-cloud-aiplatform \\\n",
" google-cloud-pipeline-components==1.0.25 \\\n",
" google-cloud-pipeline-components \\\n",
" kfp "
]
},
@@ -1,852 +0,0 @@
{
"cells": [
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "copyright"
},
"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": "title:generic,gcp"
},
"source": [
"# Get started with Model Garden Pipeline Templates for BERT models\n",
"\n",
"\n",
"<table align=\"left\">\n",
"\n",
" <td>\n",
" <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_template_pipelines_bert.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/model_garden/model_garden_template_pipelines_bert.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/communitymodel_garden/model_garden_template_pipelines_bert.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": "overview:mlops"
},
"source": [
"## Overview\n",
"\n",
"\n",
"This tutorial demonstrates how to modify, compile and execute a prebuilt Vertex AI Model Garden pipeline template with Vertex AI Pipelines.\n",
"\n",
"Learn more about [Create a pipeline template](https://cloud.google.com/vertex-ai/docs/pipelines/create-pipeline-template)."
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "objective:mlops,stage4,get_started_vertex_model_evaluation"
},
"source": [
"### Objective\n",
"\n",
"In this tutorial, you learn how to use a prebuilt pipeline template with `Vertex AI Pipelines` to fine-tune a BERT text classification model, where the model is accessed from `Vertex AI Model Garden`.\n",
"\n",
"This tutorial uses the following Google Cloud ML services:\n",
"\n",
"- `Vertex AI Pipelines`\n",
"- `Vertex AI Training`\n",
"- `Vertex AI Model Garden`\n",
"- `Google Cloud Pipeline Components`\n",
"\n",
"\n",
"The steps performed include:\n",
"\n",
"- Create a user-defined repository in the `Artifact Registry`.\n",
"- Upload the prebuilt pipeline template to the `Artifact Registry`.\n",
"- Create a pipeline job with the prebuilt pipeline template to fine-tune a BERT model.\n",
"- Execute the pipeline using `Vertex AI Pipelines`.\n",
" - Load BERT model from Vertex AI Model Garden\n",
" - Fine-tune train the model\n",
" - Do batch prediction\n",
" - Evaluate the model from the batch prediction results\n",
"- Obtain the Vertex AI Model resource from the pipeline artifacts.\n",
"- Deploy the model to a Vertex AI Endpoint\n",
"- Make a prediction"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "dataset:bank,lbn"
},
"source": [
"### Model\n",
"\n",
"This tutorial uses a pre-trained BERT text classification model from `Vertex AI Model Garden`, which is then fine-tuned (transfer learning) on a dataset of text phrases which are classified as either FirstClass or SecondClass.\n",
"\n",
"Learn more about [BERT pretrained encoder model]( https://tfhub.dev/tensorflow/bert_en_uncased_L-12_H-768_A-12/3). "
]
},
{
"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",
"* Dataflow\n",
"\n",
"Learn about [Vertex AI\n",
"pricing](https://cloud.google.com/vertex-ai/pricing), [Cloud Storage\n",
"pricing](https://cloud.google.com/storage/pricing), and [Dataflow pricing](https://cloud.google.com/dataflow/pricing)\n",
"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 packages required for executing this notebook.\n",
"\n",
"*Note:* This tutorial requires KFP 2.x."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "install_mlops"
},
"outputs": [],
"source": [
"import os\n",
"\n",
"! pip3 install --upgrade google-cloud-aiplatform \\\n",
" google-cloud-pipeline-components \\\n",
" kfp==2.0.0b15"
]
},
{
"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"
},
"source": [
"## Before you begin\n"
]
},
{
"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": "c4ccf556d4ea"
},
"source": [
"### Enable APIs\n",
"\n",
"You can enable the required APIs using `gcloud`."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "619529337e6d"
},
"outputs": [],
"source": [
"! gcloud services enable compute.googleapis.com \\\n",
" containerregistry.googleapis.com \\\n",
" aiplatform.googleapis.com \\\n",
" artifactregistry.googleapis.com"
]
},
{
"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": "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": [
"IS_COLAB = False\n",
"# from google.colab import auth\n",
"# auth.authenticate_user()\n",
"# IS_COLAB=True"
]
},
{
"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-{PROJECT_ID}-unique\" # @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": "set_service_account"
},
"source": [
"#### Service Account\n",
"\n",
"**If you don't know your service account**, try to get your service account using `gcloud` command by executing the second cell below."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "set_service_account"
},
"outputs": [],
"source": [
"SERVICE_ACCOUNT = \"[your-service-account]\" # @param {type:\"string\"}"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "autoset_service_account"
},
"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",
" if 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": "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 \n",
"! gsutil iam ch serviceAccount:{SERVICE_ACCOUNT}:roles/storage.objectViewer "
]
},
{
"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_kfp"
},
"outputs": [],
"source": [
"import json\n",
"\n",
"import google.cloud.aiplatform as aiplatform\n",
"from kfp.registry import RegistryClient"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "init_aip:mbsdk,all"
},
"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,all"
},
"outputs": [],
"source": [
"aiplatform.init(project=PROJECT_ID, location=REGION, staging_bucket=BUCKET_URI)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "9b773e8d2bd2"
},
"source": [
"## Create repo in Artifact Registry\n",
"\n",
"First, you create your own (user-defined) repository in the `Artifact Registry`. You use this repository to upload and retrieve your pipeline templates.\n",
"\n",
"The name of your repo is `quickstart-kfp-repo`"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "520de849cee2"
},
"outputs": [],
"source": [
"REPO_NAME = \"quickstart-kfp-repo\"\n",
"\n",
"! gcloud artifacts repositories create {REPO_NAME} --location={REGION} --repository-format=KFP"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "1611d3517c0f"
},
"source": [
"### Upload the pipeline template\n",
"\n",
"Next, you instantiate a client interface to the Artifact Registry. Then with the `upload_pipeline()` method you upload your pipeline template."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "72db37f6d67c"
},
"outputs": [],
"source": [
"BERT_YAML = \"gs://cloud-samples-data/vertex-ai/dataset-management/datasets/bert_finetuning/pipeline.yaml\"\n",
"\n",
"! gsutil cp {BERT_YAML} pipeline.yaml\n",
"\n",
"client = RegistryClient(\n",
" host=f\"https://{REGION}-kfp.pkg.dev/{PROJECT_ID}/quickstart-kfp-repo\"\n",
")\n",
"\n",
"templateName, versionName = client.upload_pipeline(\n",
" file_name=\"pipeline.yaml\",\n",
" tags=[\"v1\", \"latest\"],\n",
" extra_headers={\n",
" \"description\": \"This is a pipeline template for fine-tuning a BERT model.\"\n",
" },\n",
")\n",
"\n",
"! rm pipeline.yaml"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "02f82754fc0d"
},
"source": [
"### View your artifacts in your registry\n",
"\n",
"Next, using the `gcloud artifacts files` command you view the artifacts, inclusive of the pipeline template, in your artifacts repository."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "b2f641eb2056"
},
"outputs": [],
"source": [
"! gcloud artifacts files list --repository={REPO_NAME} --location={REGION}"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "9d5296831cfb"
},
"source": [
"## Load and execute the pipeline job\n",
"\n",
"Next, you create a Vertex AI Pipeline job from your BERT pipeline template by instantiating a PipelineJob(), with the following parameters:\n",
"\n",
"- `display_name`: The human readable name for the pipeline job.\n",
"- `template_path`: The path to the pipeline template in the Artifact Registry.\n",
"- `enable_caching`: On re-runs, use the results from previous successful and unchanged steps.\n",
"- `pipeline_root`: A Cloud storage location for storing pipeline results.\n",
"- `parameter_values`: The parameters and values that are input to the template pipeline. In this example, they are:\n",
" - `project`: Your project ID.\n",
" - `class_labels`: A list of valid class labels, in cardinal order.\n",
" - `root_dir`: A Cloud Storage scratch area.\n",
" - `training_data_path`: A Cloud Storage location to the training data.\n",
" - `ground_truth_gcs_source_uris`: A Cloud Storage location to evaluation data."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "a3c502fc7e41"
},
"outputs": [],
"source": [
"PIPELINE_ROOT = f\"{BUCKET_URI}/pipeline_root/bert-finetuning\"\n",
"\n",
"job = aiplatform.PipelineJob(\n",
" display_name=\"bert-finetuning\",\n",
" template_path=f\"https://{REGION}-kfp.pkg.dev/{PROJECT_ID}/quickstart-kfp-repo/{templateName}/{versionName}\",\n",
" pipeline_root=PIPELINE_ROOT,\n",
" enable_caching=False,\n",
" parameter_values={\n",
" \"project\": PROJECT_ID,\n",
" \"class_labels\": [\"FirstClass\", \"SecondClass\", \"[UNK]\"],\n",
" \"root_dir\": BUCKET_URI,\n",
" \"ground_truth_gcs_source_uris\": [\n",
" \"gs://cloud-samples-data/vertex-ai/dataset-management/datasets/bert_finetuning/wide_and_deep_trainer_container_tests_input.jsonl\"\n",
" ],\n",
" \"training_data_path\": \"gs://cloud-samples-data/vertex-ai/dataset-management/datasets/bert_finetuning/wide_and_deep_trainer_container_tests_input.jsonl\",\n",
" },\n",
")\n",
"\n",
"job.run()"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "view_pipleline_results:bqml"
},
"source": [
"### View the pipeline results"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "view_pipleline_results:bqml"
},
"outputs": [],
"source": [
"PROJECT_NUMBER = job.gca_resource.name.split(\"/\")[1]\n",
"print(PROJECT_NUMBER)\n",
"\n",
"\n",
"def print_pipeline_output(job, output_task_name):\n",
" JOB_ID = job.name\n",
" print(JOB_ID)\n",
" artifact = \"\"\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",
" EXECUTE_OUTPUT = (\n",
" PIPELINE_ROOT\n",
" + \"/\"\n",
" + PROJECT_NUMBER\n",
" + \"/\"\n",
" + JOB_ID\n",
" + \"/\"\n",
" + output_task_name\n",
" + \"_\"\n",
" + str(TASK_ID)\n",
" + \"/executor_output.json\"\n",
" )\n",
" GCP_RESOURCES = (\n",
" PIPELINE_ROOT\n",
" + \"/\"\n",
" + PROJECT_NUMBER\n",
" + \"/\"\n",
" + JOB_ID\n",
" + \"/\"\n",
" + output_task_name\n",
" + \"_\"\n",
" + str(TASK_ID)\n",
" + \"/gcp_resources\"\n",
" )\n",
" EVALUATION_METRICS = (\n",
" PIPELINE_ROOT\n",
" + \"/\"\n",
" + PROJECT_NUMBER\n",
" + \"/\"\n",
" + JOB_ID\n",
" + \"/\"\n",
" + output_task_name\n",
" + \"_\"\n",
" + str(TASK_ID)\n",
" + \"/evaluation_metrics\"\n",
" )\n",
" # Check if file exists, 0 is success\n",
" !gsutil -q stat $EXECUTE_OUTPUT\n",
" if _exit_code == 0:\n",
" ! gsutil cat $EXECUTE_OUTPUT\n",
" artifact = EXECUTE_OUTPUT\n",
" break\n",
" !gsutil -q stat $GCP_RESOURCES\n",
" if _exit_code == 0:\n",
" ! gsutil cat $GCP_RESOURCES\n",
" artifact = GCP_RESOURCES\n",
" break\n",
" !gsutil -q stat $EVALUATION_METRICS\n",
" if _exit_code == 0:\n",
" ! gsutil cat $EVALUATION_METRICS\n",
" artifact = EVALUATION_METRICS\n",
" break\n",
"\n",
" return artifact\n",
"\n",
"\n",
"print(\"get-vertex-model\")\n",
"artifacts = print_pipeline_output(job, \"get-vertex-model\")\n",
"output = !gsutil cat $artifacts\n",
"print(output)\n",
"output = json.loads(output[0])\n",
"model_id = output[\"artifacts\"][\"model\"][\"artifacts\"][0][\"metadata\"][\"resourceName\"]\n",
"print(\"\\n\\n\")"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "f431a9e6f025"
},
"source": [
"### Delete the pipeline job\n",
"\n",
"The method 'delete()' will delete the pipeline job."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "00bf554abbc6"
},
"outputs": [],
"source": [
"job.delete()"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "3d183db57ae2"
},
"source": [
"### Deploy the model\n",
"\n",
"Next, you deploy the model to an endpoint:\n",
"\n",
"- Use the `model_id` obtained from the pipeline artifacts to instaniate a Vertex AI Model resource instance.\n",
"- Deploy the Vertex AI Model resource to a Vertex AI Endpoint resource.\n"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "591ccc049ce5"
},
"outputs": [],
"source": [
"model = aiplatform.Model(model_id)\n",
"endpoint = model.deploy(\n",
" accelerator_count=1,\n",
" accelerator_type=aiplatform.gapic.AcceleratorType.NVIDIA_TESLA_T4.name,\n",
" machine_type=\"n1-standard-4\",\n",
")\n",
"print(endpoint)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "edb781a92864"
},
"source": [
"### Make a prediction\n",
"\n",
"Finally, you make a prediction with the deployed model."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "72d94012a987"
},
"outputs": [],
"source": [
"endpoint.predict([\"this is a test\"])"
]
},
{
"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",
"endpoint.undeploy_all()\n",
"endpoint.delete()\n",
"model.delete()\n",
"\n",
"if delete_bucket or os.getenv(\"IS_TESTING\"):\n",
" ! gsutil rm -r $BUCKET_URI\n",
"\n",
"! rm -rf custom custom.tar.gz"
]
}
],
"metadata": {
"colab": {
"name": "model_garden_pipeline_templates_bert.ipynb",
"toc_visible": true
},
"kernelspec": {
"display_name": "Python 3",
"name": "python3"
}
},
"nbformat": 4,
"nbformat_minor": 0
}
@@ -1,850 +0,0 @@
{
"cells": [
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "copyright"
},
"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": "title:generic,gcp"
},
"source": [
"# Get started with Model Garden Pipeline Templates for T5X models\n",
"\n",
"\n",
"<table align=\"left\">\n",
"\n",
" <td>\n",
" <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_template_pipelines_t5x.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/model_garden/model_garden_template_pipelines_t5x.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/communitymodel_garden/model_garden_template_pipelines_t5x.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": "overview:mlops"
},
"source": [
"## Overview\n",
"\n",
"\n",
"This tutorial demonstrates how to modify, compile and execute a prebuilt Vertex AI Model Garden pipeline template with Vertex AI Pipelines.\n",
"\n",
"Learn more about [Create a pipeline template](https://cloud.google.com/vertex-ai/docs/pipelines/create-pipeline-template)."
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "objective:mlops,stage4,get_started_vertex_model_evaluation"
},
"source": [
"### Objective\n",
"\n",
"In this tutorial, you learn how to use a prebuilt pipeline template with `Vertex AI Pipelines` to fine-tune a T5X text classification model, where the model is accessed from `Vertex AI Model Garden`.\n",
"\n",
"This tutorial uses the following Google Cloud ML services:\n",
"\n",
"- `Vertex AI Pipelines`\n",
"- `Vertex AI Training`\n",
"- `Vertex AI Model Garden`\n",
"- `Google Cloud Pipeline Components`\n",
"\n",
"\n",
"The steps performed include:\n",
"\n",
"- Create a user-defined repository in the `Artifact Registry`.\n",
"- Upload the prebuilt pipeline template to the `Artifact Registry`.\n",
"- Create a pipeline job with the prebuilt pipeline template to fine-tune a T5X model.\n",
"- Execute the pipeline using `Vertex AI Pipelines`.\n",
" - Load T5X model from Vertex AI Model Garden\n",
" - Fine-tune train the model\n",
"- Obtain the Vertex AI Model resource from the pipeline artifacts.\n",
"- Deploy the model to a Vertex AI Endpoint\n",
"- Make a prediction"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "dataset:bank,lbn"
},
"source": [
"### Model\n",
"\n",
"This tutorial uses a pre-trained T5 text classification model from `Vertex AI Model Garden`, which is then fine-tuned (transfer learning) on a dataset of text phrases which are classified as either FirstClass or SecondClass.\n",
"\n",
"Learn more about [Text-to-text transfer transformer](https://github.com/google-research/text-to-text-transfer-transformer). "
]
},
{
"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",
"* Dataflow\n",
"\n",
"Learn about [Vertex AI\n",
"pricing](https://cloud.google.com/vertex-ai/pricing), [Cloud Storage\n",
"pricing](https://cloud.google.com/storage/pricing), and [Dataflow pricing](https://cloud.google.com/dataflow/pricing)\n",
"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 packages required for executing this notebook.\n",
"\n",
"*Note:* This tutorial requires KFP 2.x."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "install_mlops"
},
"outputs": [],
"source": [
"import os\n",
"\n",
"! pip3 install --upgrade google-cloud-aiplatform \\\n",
" google-cloud-pipeline-components \\\n",
" kfp==2.0.0b15"
]
},
{
"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"
},
"source": [
"## Before you begin\n"
]
},
{
"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": "c4ccf556d4ea"
},
"source": [
"### Enable APIs\n",
"\n",
"You can enable the required APIs using `gcloud`."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "619529337e6d"
},
"outputs": [],
"source": [
"! gcloud services enable compute.googleapis.com \\\n",
" containerregistry.googleapis.com \\\n",
" aiplatform.googleapis.com \\\n",
" artifactregistry.googleapis.com"
]
},
{
"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": "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": [
"IS_COLAB = False\n",
"# from google.colab import auth\n",
"# auth.authenticate_user()\n",
"# IS_COLAB=True"
]
},
{
"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-{PROJECT_ID}-unique\" # @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": "set_service_account"
},
"source": [
"#### Service Account\n",
"\n",
"**If you don't know your service account**, try to get your service account using `gcloud` command by executing the second cell below."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "set_service_account"
},
"outputs": [],
"source": [
"SERVICE_ACCOUNT = \"[your-service-account]\" # @param {type:\"string\"}"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "autoset_service_account"
},
"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",
" if 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": "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 \n",
"! gsutil iam ch serviceAccount:{SERVICE_ACCOUNT}:roles/storage.objectViewer "
]
},
{
"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_kfp"
},
"outputs": [],
"source": [
"import json\n",
"\n",
"import google.cloud.aiplatform as aiplatform\n",
"from kfp.registry import RegistryClient"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "init_aip:mbsdk,all"
},
"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,all"
},
"outputs": [],
"source": [
"aiplatform.init(project=PROJECT_ID, location=REGION, staging_bucket=BUCKET_URI)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "9b773e8d2bd2"
},
"source": [
"## Create repo in Artifact Registry\n",
"\n",
"First, you create your own (user-defined) repository in the `Artifact Registry`. You use this repository to upload and retreive your pipeline templates.\n",
"\n",
"The name of your repo is `quickstart-kfp-repo`"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "520de849cee2"
},
"outputs": [],
"source": [
"REPO_NAME = \"quickstart-kfp-repo\"\n",
"\n",
"! gcloud artifacts repositories create {REPO_NAME} --location={REGION} --repository-format=KFP"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "1611d3517c0f"
},
"source": [
"### Upload the pipeline template\n",
"\n",
"Next, you instantiate a client interface to the Artifact Registry. Then with the `upload_pipeline()` method you upload your pipeline template."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "7f002e57998a"
},
"outputs": [],
"source": [
"T5X_YAML = \"gs://cloud-samples-data/vertex-ai/dataset-management/datasets/t5_finetuning/pipeline.yaml\"\n",
"\n",
"! gsutil cp {T5X_YAML} pipeline.yaml\n",
"\n",
"client = RegistryClient(\n",
" host=f\"https://{REGION}-kfp.pkg.dev/{PROJECT_ID}/quickstart-kfp-repo\"\n",
")\n",
"\n",
"templateName, versionName = client.upload_pipeline(\n",
" file_name=\"pipeline.yaml\",\n",
" tags=[\"v1\", \"latest\"],\n",
" extra_headers={\n",
" \"description\": \"This is a pipeline template for fine-tuning a T5 model.\"\n",
" },\n",
")\n",
"\n",
"! rm pipeline.yaml"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "02f82754fc0d"
},
"source": [
"### View your artifacts in your registry\n",
"\n",
"Next, using the `gcloud artifacts files` command you view the artifacts, inclusive of the pipeline template, in your artifacts repository."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "b2f641eb2056"
},
"outputs": [],
"source": [
"! gcloud artifacts files list --repository={REPO_NAME} --location={REGION}"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "968a46a3cb6d"
},
"source": [
"## Load and execute the pipeline job\n",
"\n",
"Next, you create a Vertex AI Pipeline job from your T5 pipeline template by instantiating a PipelineJob(), with the following parameters:\n",
"\n",
"- `display_name`: The human readable name for the pipeline job.\n",
"- `template_path`: The path to the pipeline template in the Artifact Registry.\n",
"- `enable_caching`: On re-runs, use the results from previous successful and unchanged steps.\n",
"- `pipeline_root`: A Cloud storage location for storing pipeline results.\n",
"- `parameter_values`: The parameters and values that are input to the template pipeline. In this example, they are:\n",
"TODO\n",
" - `project`: Your project ID.\n",
" - `class_labels`: A list of valid class labels, in cardinal order.\n",
" - `root_dir`: A Cloud Storage scratch area.\n",
" - `training_data_path`: A Cloud Storage location to the training data.\n",
" - `ground_truth_gcs_source_uris`: A Cloud Storage location to evaluation data."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "a3c502fc7e41"
},
"outputs": [],
"source": [
"PIPELINE_ROOT = f\"{BUCKET_URI}/pipeline_root/t5_finetuning\"\n",
"\n",
"job = aiplatform.PipelineJob(\n",
" display_name=\"t5x-finetuning\",\n",
" template_path=f\"https://{REGION}-kfp.pkg.dev/{PROJECT_ID}/quickstart-kfp-repo/{templateName}/{versionName}\",\n",
" pipeline_root=PIPELINE_ROOT,\n",
" enable_caching=False,\n",
" parameter_values={\n",
" \"project_id\": PROJECT_ID,\n",
" \"accelerator_count\": 32,\n",
" \"feature_keys\": \"question\",\n",
" \"label_key\": \"answer\",\n",
" \"training_data_path\": \"gs://cloud-llm-public/tfds/natural_questions_open/1.0.0_shortened/natural_questions_open-train.tfrecord-00000-of-00001\",\n",
" \"validation_data_path\": \"gs://cloud-llm-public/tfds/natural_questions_open/1.0.0_shortened/natural_questions_open-validation.tfrecord-00000-of-00001\",\n",
" },\n",
")\n",
"\n",
"job.run()"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "view_pipleline_results:bqml"
},
"source": [
"### View the pipeline results"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "view_pipleline_results:bqml"
},
"outputs": [],
"source": [
"PROJECT_NUMBER = job.gca_resource.name.split(\"/\")[1]\n",
"print(PROJECT_NUMBER)\n",
"\n",
"\n",
"def print_pipeline_output(job, output_task_name):\n",
" JOB_ID = job.name\n",
" print(JOB_ID)\n",
" artifact = \"\"\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",
" EXECUTE_OUTPUT = (\n",
" PIPELINE_ROOT\n",
" + \"/\"\n",
" + PROJECT_NUMBER\n",
" + \"/\"\n",
" + JOB_ID\n",
" + \"/\"\n",
" + output_task_name\n",
" + \"_\"\n",
" + str(TASK_ID)\n",
" + \"/executor_output.json\"\n",
" )\n",
" GCP_RESOURCES = (\n",
" PIPELINE_ROOT\n",
" + \"/\"\n",
" + PROJECT_NUMBER\n",
" + \"/\"\n",
" + JOB_ID\n",
" + \"/\"\n",
" + output_task_name\n",
" + \"_\"\n",
" + str(TASK_ID)\n",
" + \"/gcp_resources\"\n",
" )\n",
" EVALUATION_METRICS = (\n",
" PIPELINE_ROOT\n",
" + \"/\"\n",
" + PROJECT_NUMBER\n",
" + \"/\"\n",
" + JOB_ID\n",
" + \"/\"\n",
" + output_task_name\n",
" + \"_\"\n",
" + str(TASK_ID)\n",
" + \"/evaluation_metrics\"\n",
" )\n",
" # Check if file exists, 0 is success\n",
" !gsutil -q stat $EXECUTE_OUTPUT\n",
" if _exit_code == 0:\n",
" ! gsutil cat $EXECUTE_OUTPUT\n",
" artifact = EXECUTE_OUTPUT\n",
" break\n",
" !gsutil -q stat $GCP_RESOURCES\n",
" if _exit_code == 0:\n",
" ! gsutil cat $GCP_RESOURCES\n",
" artifact = GCP_RESOURCES\n",
" break\n",
" !gsutil -q stat $EVALUATION_METRICS\n",
" if _exit_code == 0:\n",
" ! gsutil cat $EVALUATION_METRICS\n",
" artifact = EVALUATION_METRICS\n",
" break\n",
"\n",
" return artifact\n",
"\n",
"\n",
"print(\"model-upload\")\n",
"artifacts = print_pipeline_output(job, \"model-upload\")\n",
"output = !gsutil cat $artifacts\n",
"print(output)\n",
"output = json.loads(output[0])\n",
"model_id = output[\"artifacts\"][\"model\"][\"artifacts\"][0][\"metadata\"][\"resourceName\"]\n",
"print(\"\\n\\n\")"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "f431a9e6f025"
},
"source": [
"### Delete the pipeline job\n",
"\n",
"The method 'delete()' will delete the pipeline job."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "00bf554abbc6"
},
"outputs": [],
"source": [
"job.delete()"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "3d183db57ae2"
},
"source": [
"### Deploy the model\n",
"\n",
"Next, you deploy the model to an endpoint:\n",
"\n",
"- Use the `model_id` obtained from the pipeline artifacts to instantiate a Vertex AI Model resource instance.\n",
"- Deploy the Vertex AI Model resource to a Vertex AI Endpoint resource.\n"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "591ccc049ce5"
},
"outputs": [],
"source": [
"model = aiplatform.Model(model_id)\n",
"endpoint = model.deploy(\n",
" accelerator_count=1,\n",
" accelerator_type=aiplatform.gapic.AcceleratorType.NVIDIA_TESLA_T4.name,\n",
" machine_type=\"n1-standard-4\",\n",
")\n",
"print(endpoint)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "edb781a92864"
},
"source": [
"### Make a prediction\n",
"\n",
"Finally, you make a prediction with the deployed model."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "72d94012a987"
},
"outputs": [],
"source": [
"endpoint.predict([\"this is a test\"])"
]
},
{
"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 = True\n",
"\n",
"endpoint.undeploy_all()\n",
"endpoint.delete()\n",
"model.delete()\n",
"\n",
"if delete_bucket or os.getenv(\"IS_TESTING\"):\n",
" ! gsutil rm -r $BUCKET_URI\n",
"\n",
"! rm -rf custom custom.tar.gz"
]
}
],
"metadata": {
"colab": {
"name": "model_garden_pipeline_templates_t5x.ipynb",
"toc_visible": true
},
"kernelspec": {
"display_name": "Python 3",
"name": "python3"
}
},
"nbformat": 4,
"nbformat_minor": 0
}
@@ -271,7 +271,7 @@
"\n",
"\n",
"def deploy_model(model_id, task):\n",
" model_name = \"blip2\"\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",
@@ -345,10 +345,10 @@
{
"cell_type": "markdown",
"metadata": {
"id": "80b3fd2ace09"
"id": "12893aa2c5af"
},
"source": [
"NOTE: The model weights will be downloaded after the deployment succeeds. Thus additional 10 minutes of waiting time is needed **after** the above model deployment step succeeds and before you run the next step below. Otherwise you might see a `ServiceUnavailable: 503 502:Bad Gateway` error when you send requests to the endpoint."
"NOTE: The model weights will be downloaded after the deployment succeeds. When the model is very large it could add 5~15mins additional time before the endpoint is ready for prediction."
]
},
{
@@ -409,10 +409,10 @@
{
"cell_type": "markdown",
"metadata": {
"id": "80b3fd2ace09"
"id": "0ac7f8d945e3"
},
"source": [
"NOTE: The model weights will be downloaded after the deployment succeeds. Thus additional 10 minutes of waiting time is needed **after** the above model deployment step succeeds and before you run the next step below. Otherwise you might see a `ServiceUnavailable: 503 502:Bad Gateway` error when you send requests to the endpoint."
"NOTE: The model weights will be downloaded after the deployment succeeds. When the model is very large it could add 5~15mins additional time before the endpoint is ready for prediction."
]
},
{
@@ -332,15 +332,6 @@
")"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "80b3fd2ace09"
},
"source": [
"NOTE: The model weights will be downloaded after the deployment succeeds. Thus additional 5 minutes of waiting time is needed **after** the above model deployment step succeeds and before you run the next step below. Otherwise you might see a `ServiceUnavailable: 503 502:Bad Gateway` error when you send requests to the endpoint."
]
},
{
"cell_type": "code",
"execution_count": null,
@@ -332,15 +332,6 @@
")"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "80b3fd2ace09"
},
"source": [
"NOTE: The model weights will be downloaded after the deployment succeeds. Thus additional 5 minutes of waiting time is needed **after** the above model deployment step succeeds and before you run the next step below. Otherwise you might see a `ServiceUnavailable: 503 502:Bad Gateway` error when you send requests to the endpoint."
]
},
{
"cell_type": "code",
"execution_count": null,
@@ -314,16 +314,9 @@
"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."
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "99f3c6b404b5"
},
"source": [
"### Zero-shot image classification"
"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."
]
},
{
@@ -339,15 +332,6 @@
")"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "80b3fd2ace09"
},
"source": [
"NOTE: The model weights will be downloaded after the deployment succeeds. Thus additional 5 minutes of waiting time is needed **after** the above model deployment step succeeds and before you run the next step below. Otherwise you might see a `ServiceUnavailable: 503 502:Bad Gateway` error when you send requests to the endpoint."
]
},
{
"cell_type": "code",
"execution_count": null,
@@ -369,6 +353,15 @@
"print(preds)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "db7ffebdb4be"
},
"source": [
"### Clean up resources"
]
},
{
"cell_type": "code",
"execution_count": null,
@@ -383,82 +376,6 @@
"# Delete models.\n",
"model.delete()"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "ddf9e9ec7b58"
},
"source": [
"### Image/text feature embedding"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "8897bbac6887"
},
"outputs": [],
"source": [
"model, endpoint = deploy_model(\n",
" model_id=\"openai/clip-vit-base-patch32\", task=\"feature-embedding\"\n",
")"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "9f854c45f8f3"
},
"source": [
"NOTE: The model weights will be downloaded after the deployment succeeds. Thus additional 5 minutes of waiting time is needed after the above model deployment step succeeds and before you run the next step below. Otherwise you might see a ServiceUnavailable: 503 502:Bad Gateway error when you send requests to the endpoint."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "2abd54335f36"
},
"outputs": [],
"source": [
"import numpy as np\n",
"\n",
"# Extract feature embedding of images.\n",
"image = download_image(\"http://images.cocodataset.org/val2017/000000039769.jpg\")\n",
"display(image)\n",
"instances = [\n",
" {\"image\": image_to_base64(image)},\n",
"]\n",
"preds = endpoint.predict(instances=instances).predictions\n",
"image_features = np.array(preds[0][\"image_features\"])\n",
"print(image_features.shape)\n",
"\n",
"# Extract feature embedding of texts.\n",
"instances = [\n",
" {\"text\": \"two cats\"},\n",
" {\"text\": \"hello world\"},\n",
"]\n",
"preds = endpoint.predict(instances=instances).predictions\n",
"text_features = np.array(preds[0][\"text_features\"])\n",
"print(text_features.shape)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "712eb9d0b336"
},
"outputs": [],
"source": [
"# Undeploy model and delete endpoint.\n",
"endpoint.delete(force=True)\n",
"\n",
"# Delete models.\n",
"model.delete()"
]
}
],
"metadata": {
@@ -184,9 +184,7 @@
"\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.\n",
"\n",
"1. [Create a service account](https://cloud.google.com/iam/docs/service-accounts-create#iam-service-accounts-create-console) with `Vertex AI User` and `Storage Object Admin` roles for deploying fine tuned model to Vertex AI endpoint."
"1. [Create a Cloud Storage bucket](https://cloud.google.com/storage/docs/creating-buckets) for storing experiment outputs."
]
},
{
@@ -213,10 +211,7 @@
"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\"}\n",
"\n",
"# The service account you created in step-5 above, it's like \"<account_name>@<project>.iam.gserviceaccount.com\"\n",
"SERVICE_ACCOUNT = \"\" # @param {type:\"string\"}"
"GCS_BUCKET = \"\" # @param {type:\"string\"}"
]
},
{
@@ -276,15 +271,13 @@
},
{
"cell_type": "code",
"execution_count": null,
"execution_count": 1,
"metadata": {
"id": "354da31189dc"
},
"outputs": [],
"source": [
"import base64\n",
"import os\n",
"from datetime import datetime\n",
"from io import BytesIO\n",
"\n",
"import cv2\n",
@@ -341,6 +334,8 @@
" \"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",
@@ -348,6 +343,7 @@
" 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",
@@ -355,7 +351,6 @@
" accelerator_type=\"NVIDIA_TESLA_V100\",\n",
" accelerator_count=1,\n",
" deploy_request_timeout=1800,\n",
" service_account=SERVICE_ACCOUNT,\n",
" )\n",
" return model, endpoint"
]
@@ -415,13 +410,6 @@
"# for a full list of training arguments.\n",
"model = job.run(\n",
" args=[\n",
" f\"--num_machines={num_nodes}\",\n",
" f\"--num_processes={num_gpus}\",\n",
" \"--machine_rank=0\",\n",
" \"--mixed_precision=no\",\n",
" \"--gpu_ids=all\",\n",
" \"--same_network\",\n",
" \"--dynamo_backend=no\",\n",
" \"controlnet/train_controlnet.py\",\n",
" \"--tracker_project_name=train_controlnet\",\n",
" f\"--pretrained_model_name_or_path={stable_diffusion_model_id}\",\n",
@@ -430,8 +418,6 @@
" \"--resolution=512\",\n",
" \"--learning_rate=1e-5\",\n",
" \"--train_batch_size=2\",\n",
" \"--checkpointing_steps=50000\",\n",
" \"--checkpoints_total_limit=1\",\n",
" ],\n",
" replica_count=num_nodes,\n",
" machine_type=machine_type,\n",
@@ -491,15 +477,6 @@
")"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "80b3fd2ace09"
},
"source": [
"NOTE: The model weights will be downloaded after the deployment succeeds. Thus additional 5 minutes of waiting time is needed **after** the above model deployment step succeeds and before you run the next step below. Otherwise you might see a `ServiceUnavailable: 503 502:Bad Gateway` error when you send requests to the endpoint."
]
},
{
"cell_type": "code",
"execution_count": null,
@@ -543,11 +520,11 @@
},
"outputs": [],
"source": [
"# Undeploy model and delete endpoint.\n",
"endpoint.delete(force=True)\n",
"\n",
"# Delete models.\n",
"model.delete()"
"model.delete()\n",
"\n",
"# Undeploy model and delete endpoint.\n",
"endpoint.delete(force=True)"
]
},
{
@@ -577,19 +554,10 @@
"outputs": [],
"source": [
"model, endpoint = deploy_model(\n",
" model_id=f\"gs://{GCS_BUCKET}/controlnet/output\", task=\"controlnet\"\n",
" model_id=f\"gs://{GCS_BUCKET}/controlnet/output\", task=\"image-to-image\"\n",
")"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "80b3fd2ace09"
},
"source": [
"NOTE: The model weights will be downloaded after the deployment succeeds. Thus additional 5 minutes of waiting time is needed **after** the above model deployment step succeeds and before you run the next step below. Otherwise you might see a `ServiceUnavailable: 503 502:Bad Gateway` error when you send requests to the endpoint."
]
},
{
"cell_type": "code",
"execution_count": null,
File diff suppressed because it is too large Load Diff
@@ -352,15 +352,6 @@
")"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "80b3fd2ace09"
},
"source": [
"NOTE: The model weights will be downloaded after the deployment succeeds. Thus additional 5 minutes of waiting time is needed **after** the above model deployment step succeeds and before you run the next step below. Otherwise you might see a `ServiceUnavailable: 503 502:Bad Gateway` error when you send requests to the endpoint."
]
},
{
"cell_type": "code",
"execution_count": null,
@@ -332,15 +332,6 @@
")"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "80b3fd2ace09"
},
"source": [
"NOTE: The model weights will be downloaded after the deployment succeeds. Thus additional 5 minutes of waiting time is needed **after** the above model deployment step succeeds and before you run the next step below. Otherwise you might see a `ServiceUnavailable: 503 502:Bad Gateway` error when you send requests to the endpoint."
]
},
{
"cell_type": "code",
"execution_count": null,
@@ -351,15 +351,6 @@
")"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "80b3fd2ace09"
},
"source": [
"NOTE: The model weights will be downloaded after the deployment succeeds. Thus additional 5 minutes of waiting time is needed **after** the above model deployment step succeeds and before you run the next step below. Otherwise you might see a `ServiceUnavailable: 503 502:Bad Gateway` error when you send requests to the endpoint."
]
},
{
"cell_type": "code",
"execution_count": null,
@@ -24,7 +24,6 @@
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "99c1c3fc2ca5"
@@ -49,13 +48,12 @@
" <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",
" (a Python-3 GPU notebook with preinstalled HuggingFace/transformer libraries is recommended)\n",
" (a Python-3 CPU notebook is recommended)\n",
" </td>\n",
"</table>"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "3de7470326a2"
@@ -63,15 +61,13 @@
"source": [
"## Overview\n",
"\n",
"This notebook demonstrates running local inference on [Vertex AI Workbench](https://cloud.google.com/vertex-ai-workbench).\n",
"This notebook also 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",
"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",
"- Run local predictions for text-to-image and text-guided-image-to-image with serving dockers.\n",
"- Finetune the stable-diffusion-v1.5 model with [Dreambooth](https://huggingface.co/docs/diffusers/training/dreambooth).\n",
"- Upload the model to [Vertex AI Model Registry](https://cloud.google.com/vertex-ai/docs/model-registry/introduction).\n",
"- Deploy the model to a [Vertex AI Endpoint resource](https://cloud.google.com/vertex-ai/docs/predictions/using-private-endpoints).\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",
@@ -85,114 +81,6 @@
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "e8a42fa49305"
},
"source": [
"## Local inference (Workbench only)"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "1169c41b76b3"
},
"source": [
"The quickest and easiest way to use this model locally is by using Vertex AI Workbench with a pre-built custom container that has the necessary packages installed.\n",
"\n",
"\n",
"\n",
"1. Follow [this link](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_pytorch_stable_diffusion.ipynb) to deploy the notebook to a Vertex AI Workbench Instance.\n",
"2. Select `Create a new Notebook`.\n",
"3. Click `Advanced Options`.\n",
"4. Under **Environment**, select `Custom Container` for `Environment`. \n",
"5. Set `Docker container image` to `us-docker.pkg.dev/vertex-ai/vertex-vision-model-garden-dockers/transformers-notebook`.\n",
"6. Under **Machine configuration**, select a GPU and select `Install NVIDIA GPU driver automatically for me`.\n",
"7. Click `Create` to create the Vertex AI Workbench instance. \n",
"\n",
"Once the notebook is ready, simply execute the code block(s) below in the Workbench instance."
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "1d5ebc91c786"
},
"source": [
"### Text-to-image"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "d39ed8c97cc5"
},
"outputs": [],
"source": [
"import torch\n",
"from diffusers import StableDiffusionPipeline\n",
"\n",
"model_id = \"runwayml/stable-diffusion-v1-5\"\n",
"pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)\n",
"pipe = pipe.to(\"cuda\")\n",
"\n",
"prompt = \"a photo of an astronaut riding a horse on mars\"\n",
"image = pipe(prompt).images[0]\n",
"\n",
"display(image)"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "5aed5ed7b6f6"
},
"source": [
"### Text-guided image-to-image"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "faadabb7728f"
},
"outputs": [],
"source": [
"from io import BytesIO\n",
"\n",
"import requests\n",
"import torch\n",
"from diffusers import StableDiffusionImg2ImgPipeline\n",
"from PIL import Image\n",
"\n",
"device = \"cuda\"\n",
"model_id_or_path = \"runwayml/stable-diffusion-v1-5\"\n",
"pipe = StableDiffusionImg2ImgPipeline.from_pretrained(\n",
" model_id_or_path, torch_dtype=torch.float16\n",
")\n",
"pipe = pipe.to(device)\n",
"\n",
"url = \"https://raw.githubusercontent.com/CompVis/stable-diffusion/main/assets/stable-samples/img2img/sketch-mountains-input.jpg\"\n",
"\n",
"response = requests.get(url)\n",
"init_image = Image.open(BytesIO(response.content)).convert(\"RGB\")\n",
"init_image = init_image.resize((768, 512))\n",
"\n",
"prompt = \"A fantasy landscape, trending on artstation\"\n",
"\n",
"images = pipe(prompt=prompt, image=init_image, strength=0.75, guidance_scale=7.5).images\n",
"display(images[0])"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "264c07757582"
@@ -204,7 +92,6 @@
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "d73ffa0c0b83"
@@ -238,7 +125,6 @@
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "fb671e75ca7b"
@@ -256,11 +142,12 @@
"outputs": [],
"source": [
"# Install gdown for downloading example training images.\n",
"!pip install gdown"
"!pip install gdown\n",
"# Install gsutil for downloading/uploading data from/to Cloud Storage buckets.\n",
"!pip install gsutil"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "5244aac3d929"
@@ -284,7 +171,6 @@
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "bb7adab99e41"
@@ -298,13 +184,10 @@
"\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.\n",
"\n",
"1. [Create a service account](https://cloud.google.com/iam/docs/service-accounts-create#iam-service-accounts-create-console) with `Vertex AI User` and `Storage Object Admin` roles for deploying fine tuned model to Vertex AI endpoint."
"1. [Create a Cloud Storage bucket](https://cloud.google.com/storage/docs/creating-buckets) for storing experiment outputs."
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "6c460088b873"
@@ -328,14 +211,10 @@
"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\"}\n",
"\n",
"# The service account for deploying fine tuned model.\n",
"SERVICE_ACCOUNT = \"\" # @param {type:\"string\"}"
"GCS_BUCKET = \"\" # @param {type:\"string\"}"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "e828eb320337"
@@ -358,7 +237,6 @@
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "2cc825514deb"
@@ -383,7 +261,6 @@
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "0c250872074f"
@@ -401,13 +278,12 @@
"outputs": [],
"source": [
"import base64\n",
"import glob\n",
"import os\n",
"from datetime import datetime\n",
"from io import BytesIO\n",
"\n",
"import requests\n",
"from google.cloud import aiplatform, storage\n",
"from google.cloud import aiplatform\n",
"from PIL import Image\n",
"\n",
"\n",
@@ -450,6 +326,8 @@
" \"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",
@@ -457,6 +335,7 @@
" 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",
@@ -464,35 +343,11 @@
" accelerator_type=\"NVIDIA_TESLA_V100\",\n",
" accelerator_count=1,\n",
" deploy_request_timeout=1800,\n",
" service_account=SERVICE_ACCOUNT,\n",
" )\n",
" return model, endpoint\n",
"\n",
"\n",
"def get_bucket_and_blob_name(filepath):\n",
" # The gcs path is of the form gs://<bucket-name>/<blob-name>\n",
" gs_suffix = filepath.split(\"gs://\", 1)[1]\n",
" return tuple(gs_suffix.split(\"/\", 1))\n",
"\n",
"\n",
"def upload_local_dir_to_gcs(local_dir_path, gcs_dir_path):\n",
" \"\"\"Uploads files in a local directory to a GCS directory.\"\"\"\n",
" client = storage.Client()\n",
" bucket_name = gcs_dir_path.split(\"/\")[2]\n",
" bucket = client.get_bucket(bucket_name)\n",
" for local_file in glob.glob(local_dir_path + \"/**\"):\n",
" if not os.path.isfile(local_file):\n",
" continue\n",
" filename = local_file[1 + len(local_dir_path) :]\n",
" gcs_file_path = os.path.join(gcs_dir_path, filename)\n",
" _, blob_name = get_bucket_and_blob_name(gcs_file_path)\n",
" blob = bucket.blob(blob_name)\n",
" blob.upload_from_filename(local_file)\n",
" print(\"Copied {} to {}.\".format(local_file, gcs_file_path))"
" return model, endpoint"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "e70e3519ff8b"
@@ -502,7 +357,6 @@
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "0dc65d8f0689"
@@ -527,12 +381,11 @@
"!gdown --folder https://drive.google.com/drive/folders/1BO_dyz-p65qhBRRMRA4TbZ8qW4rB99JZ\n",
"\n",
"# Upload data to Cloud Storage bucket.\n",
"upload_local_dir_to_gcs(\"dog\", f\"gs://{GCS_BUCKET}/dreambooth/dog\")\n",
"upload_local_dir_to_gcs(\"dog\", f\"gs://{GCS_BUCKET}/dreambooth/dog_class\")"
"!gsutil -m cp -r dog/* gs://{GCS_BUCKET}/dreambooth/dog/\n",
"!gsutil -m cp -r dog/* gs://{GCS_BUCKET}/dreambooth/dog_class/"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "969cfeb79317"
@@ -602,7 +455,6 @@
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "bf7f82732e61"
@@ -612,7 +464,6 @@
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "1cc26e68d7b0"
@@ -624,7 +475,6 @@
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "cd7b56421392"
@@ -634,7 +484,6 @@
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "6d331b1ea337"
@@ -655,22 +504,12 @@
},
"outputs": [],
"source": [
"# Set the model_id to \"runwayml/stable-diffusion-v1-5\" to load the OSS pre-trained model.\n",
"# 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=f\"gs://{GCS_BUCKET}/dreambooth/output\", task=\"text-to-image\"\n",
" model_id=\"runwayml/stable-diffusion-v1-5\", task=\"text-to-image\"\n",
")"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "80b3fd2ace09"
},
"source": [
"NOTE: The model weights will be downloaded after the deployment succeeds. Thus additional 5 minutes of waiting time is needed **after** the above model deployment step succeeds and before you run the next step below. Otherwise you might see a `ServiceUnavailable: 503 502:Bad Gateway` error when you send requests to the endpoint."
]
},
{
"cell_type": "code",
"execution_count": null,
@@ -691,7 +530,6 @@
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "af21a3cff1e0"
@@ -716,7 +554,6 @@
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "c1e51f764a60"
@@ -726,7 +563,6 @@
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "fa686a54047c"
@@ -749,16 +585,6 @@
")"
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "80b3fd2ace09"
},
"source": [
"NOTE: The model weights will be downloaded after the deployment succeeds. Thus additional 5 minutes of waiting time is needed **after** the above model deployment step succeeds and before you run the next step below. Otherwise you might see a `ServiceUnavailable: 503 502:Bad Gateway` error when you send requests to the endpoint."
]
},
{
"cell_type": "code",
"execution_count": null,
@@ -783,7 +609,6 @@
]
},
{
"attachments": {},
"cell_type": "markdown",
"metadata": {
"id": "ed3795d474b9"
@@ -142,7 +142,9 @@
"outputs": [],
"source": [
"# Install gdown for downloading example training images.\n",
"!pip install gdown"
"!pip install gdown\n",
"# Install gsutil for downloading/uploading data from/to Cloud Storage buckets.\n",
"!pip install gsutil"
]
},
{
@@ -182,9 +184,7 @@
"\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.\n",
"\n",
"1. [Create a service account](https://cloud.google.com/iam/docs/service-accounts-create#iam-service-accounts-create-console) with `Vertex AI User` and `Storage Object Admin` roles for deploying fine tuned model to Vertex AI endpoint."
"1. [Create a Cloud Storage bucket](https://cloud.google.com/storage/docs/creating-buckets) for storing experiment outputs."
]
},
{
@@ -210,14 +210,8 @@
"# 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.\n",
"# Fill it without the 'gs://' prefix.\n",
"GCS_BUCKET = \"\" # @param {type:\"string\"}\n",
"\n",
"# The service account for deploying fine tuned model.\n",
"# The service account looks like:\n",
"# '<account_name>@<project>.iam.gserviceaccount.com'\n",
"SERVICE_ACCOUNT = \"\" # @param {type:\"string\"}"
"# The Cloud Storage bucket for storing experiments output. Fill it without the 'gs://' prefix.\n",
"GCS_BUCKET = \"\" # @param {type:\"string\"}"
]
},
{
@@ -277,20 +271,19 @@
},
{
"cell_type": "code",
"execution_count": null,
"execution_count": 2,
"metadata": {
"id": "8759e624ebc0"
},
"outputs": [],
"source": [
"import base64\n",
"import glob\n",
"import os\n",
"from datetime import datetime\n",
"from io import BytesIO\n",
"\n",
"import requests\n",
"from google.cloud import aiplatform, storage\n",
"from google.cloud import aiplatform\n",
"from PIL import Image\n",
"\n",
"\n",
@@ -333,6 +326,8 @@
" \"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",
@@ -340,6 +335,7 @@
" 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",
@@ -347,31 +343,8 @@
" accelerator_type=\"NVIDIA_TESLA_V100\",\n",
" accelerator_count=1,\n",
" deploy_request_timeout=1800,\n",
" service_account=SERVICE_ACCOUNT,\n",
" )\n",
" return model, endpoint\n",
"\n",
"\n",
"def get_bucket_and_blob_name(filepath):\n",
" # The gcs path is of the form gs://<bucket-name>/<blob-name>\n",
" gs_suffix = filepath.split(\"gs://\", 1)[1]\n",
" return tuple(gs_suffix.split(\"/\", 1))\n",
"\n",
"\n",
"def upload_local_dir_to_gcs(local_dir_path, gcs_dir_path):\n",
" \"\"\"Uploads files in a local directory to a GCS directory.\"\"\"\n",
" client = storage.Client()\n",
" bucket_name = gcs_dir_path.split(\"/\")[2]\n",
" bucket = client.get_bucket(bucket_name)\n",
" for local_file in glob.glob(local_dir_path + \"/**\"):\n",
" if not os.path.isfile(local_file):\n",
" continue\n",
" filename = local_file[1 + len(local_dir_path) :]\n",
" gcs_file_path = os.path.join(gcs_dir_path, filename)\n",
" _, blob_name = get_bucket_and_blob_name(gcs_file_path)\n",
" blob = bucket.blob(blob_name)\n",
" blob.upload_from_filename(local_file)\n",
" print(\"Copied {} to {}.\".format(local_file, gcs_file_path))"
" return model, endpoint"
]
},
{
@@ -408,8 +381,17 @@
"!gdown --folder https://drive.google.com/drive/folders/1BO_dyz-p65qhBRRMRA4TbZ8qW4rB99JZ\n",
"\n",
"# Upload data to Cloud Storage bucket.\n",
"upload_local_dir_to_gcs(\"dog\", f\"gs://{GCS_BUCKET}/dreambooth/dog\")\n",
"upload_local_dir_to_gcs(\"dog\", f\"gs://{GCS_BUCKET}/dreambooth/dog_class\")"
"!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."
]
},
{
@@ -522,21 +504,12 @@
},
"outputs": [],
"source": [
"# Set the model_id to \"runwayml/stable-diffusion-inpainting\" to load the OSS pre-trained model.\n",
"# 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=f\"gs://{GCS_BUCKET}/dreambooth/output\", task=\"image-inpainting\"\n",
" model_id=\"runwayml/stable-diffusion-inpainting\", task=\"image-inpainting\"\n",
")"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "80b3fd2ace09"
},
"source": [
"NOTE: The model weights will be downloaded after the deployment succeeds. Thus additional 5 minutes of waiting time is needed **after** the above model deployment step succeeds and before you run the next step below. Otherwise you might see a `ServiceUnavailable: 503 502:Bad Gateway` error when you send requests to the endpoint."
]
},
{
"cell_type": "code",
"execution_count": null,
@@ -1,559 +0,0 @@
{
"cells": [
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "ApNHZJmT2AMH"
},
"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": "YfxrFG052AMI"
},
"source": [
"# Vertex AI Model Garden - TIMM\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_timm.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/model_garden/model_garden_pytorch_timm.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/notebooks/deploy-notebook?download_url=https://raw.githubusercontent.com/GoogleCloudPlatform/vertex-ai-samples/main/notebooks/community/model_garden/model_garden_pytorch_timm.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",
" (a Python-3 CPU notebook is recommended)\n",
" </td>\n",
"</table>"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "76BCoQcm2AMJ"
},
"source": [
"## Overview\n",
"\n",
"This notebook demonstrates finetuning the PyTorch [timm](https://github.com/rwightman/pytorch-image-models) models and deploying the models on [Vertex AI](https://cloud.google.com/vertex-ai).\n",
"\n",
"### Objective\n",
"\n",
"- Setup environment.\n",
"- Create a custom training job on Vertex AI to train or finetune a model.\n",
"- Deploy the model on Vertex AI for online prediction.\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": "9iU1NKfh2AMJ"
},
"source": [
"## Setup environment"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "7l3GN-QL2AMJ"
},
"source": [
"### Setup 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). 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.\n",
"\n",
"1. [Enable Artifact Registry](https://cloud.google.com/artifact-registry/docs/enable-service) and [create a repository](https://cloud.google.com/artifact-registry/docs/repositories/create-repos) for storing docker images.\n",
"\n",
"1. [Create a Cloud Storage bucket](https://cloud.google.com/storage/docs/creating-buckets) for storing experiment outputs.\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 service account](https://cloud.google.com/iam/docs/service-accounts-create#iam-service-accounts-create-console) with `Vertex AI User` and `Storage Object Admin` roles for deploying fine tuned model to Vertex AI endpoint."
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "hyAQXmgf2AMK"
},
"source": [
"### Setup required libraries\n",
"\n",
"It's highly recommended to run this notebook on [Vertex AI workbench](https://cloud.google.com/vertex-ai-workbench), where you don't need to manually install any additional libraries.\n",
"\n",
"If you are running this notebook locally, you will need to install the [Cloud SDK](https://cloud.google.com/sdk) and [gsutil](https://cloud.google.com/storage/docs/gsutil_install)."
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "-uF7Kb112AMK"
},
"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": "2lP6dnfy2AMK"
},
"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": "fo65Wg9Y2AMK"
},
"source": [
"### Setup environment variables\n",
"\n",
"This notebook supports models in https://huggingface.co/docs/timm/models.\n",
"\n",
"You can also run\n",
"`python -c \"from timm import list_models; print(list_models(pretrained=True))\"`\n",
"locally to see all pretrained models.\n",
"\n",
"The following models have been manually verified to work with this notebook:\n",
"\n",
"* vit_tiny_patch16_224\n",
"* beit_base_patch16_224\n",
"* deit3_small_patch16_224\n",
"* efficientnet_b2\n",
"* mobilenetv2_100\n",
"* resnet50\n",
"* resnest50d\n",
"* convnext_base\n",
"* cspdarknet53\n",
"* inception_v4"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "3msH2i5V2AMK"
},
"outputs": [],
"source": [
"# The cloud project id.\n",
"PROJECT_ID = \"\" # @param {type:\"string\"}\n",
"# The region for running jobs.\n",
"REGION = \"us-central1\" # @param {type:\"string\"}\n",
"\n",
"# The model you want to train and serve. Please select a model from the verified model list above.\n",
"# We use a ViT model as the example.\n",
"MODEL_NAME = \"vit_tiny_patch16_224\" # @param {type:\"string\"}\n",
"\n",
"# The Cloud Storage bucket name without gs:// prefix for training outputs.\n",
"# For example: test_bucket\n",
"GCS_BUCKET = \"\" # @param {type:\"string\"}\n",
"\n",
"# The service account for deploying fine tuned model. It looks like:\n",
"# '<account_name>@<project>.iam.gserviceaccount.com'\n",
"# Follow step 6 above to create this account.\n",
"SERVICE_ACCOUNT = \"\" # @param {type:\"string\"}"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "JlutYfxH2AMK"
},
"source": [
"## Run training jobs\n",
"\n",
"This section runs a regular training job or a hyperparameter tuning job on Vertex AI.\n",
"\n",
"Before creating a training job, you need to prepare the dataset for training and evaluation.\n",
"\n",
"For example, you can use [ImageNet-1K](https://huggingface.co/datasets/imagenet-1k) held on a Cloud Storage bucket as the input dataset."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "qy4gKksX2AMK"
},
"outputs": [],
"source": [
"# The prebuilt training docker uri.\n",
"TRAIN_DOCKER_URI = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/pytorch-timm-train\"\n",
"\n",
"# The path to data directory on Cloud Storage without gs:// prefix.\n",
"# In the form of: <bucket-name>/path-to-data\n",
"GCS_DATA_DIR = \"\" # @param {type:\"string\"}"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "0DnlJhO72AMK"
},
"source": [
"### Create a training job on Vertex AI\n",
"\n",
"This section creates a training job on Vertex AI. If you want to create a hyperparameter tuning job instead, you can skip to the next section."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "1R5JWJGS2AML"
},
"outputs": [],
"source": [
"from google.cloud import aiplatform\n",
"\n",
"# Init common setup.\n",
"aiplatform.init(project=PROJECT_ID, location=REGION, staging_bucket=GCS_BUCKET)\n",
"\n",
"# Input and output path.\n",
"data_dir = f\"/gcs/{GCS_DATA_DIR}\"\n",
"output_dir = f\"/gcs/{GCS_BUCKET}/timm\"\n",
"\n",
"# Worker pool spec.\n",
"# Single node with multiple GPUs.\n",
"machine_type = \"n1-highmem-32\"\n",
"num_nodes = 1\n",
"gpu_type = \"NVIDIA_TESLA_P100\" # @param {type:\"string\"}\n",
"num_gpus = 4 # @param {type:\"integer\"}\n",
"\n",
"# Model specific config.\n",
"job_name = f\"pytorch-{MODEL_NAME}\"\n",
"batch_size = 32\n",
"epochs = 2\n",
"\n",
"job = aiplatform.CustomContainerTrainingJob(\n",
" display_name=job_name,\n",
" container_uri=TRAIN_DOCKER_URI,\n",
")\n",
"model = job.run(\n",
" args=[\n",
" \"--standalone\",\n",
" f\"--nnodes={num_nodes}\",\n",
" f\"--nproc_per_node={num_gpus}\",\n",
" \"train.py\",\n",
" data_dir,\n",
" f\"--model={MODEL_NAME}\",\n",
" \"--pretrained\",\n",
" f\"--output={output_dir}\",\n",
" f\"--batch-size={batch_size}\",\n",
" f\"--epochs={epochs}\",\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": "Tf6P7ZI82AML"
},
"source": [
"### Create a hyperparameter tuning job on Vertex AI\n",
"\n",
"You can use a [hyperparameter tuning](https://cloud.google.com/vertex-ai/docs/training/hyperparameter-tuning-overview) job to find the best configuration of your hyperparameters.\n",
"\n",
"You can skip this section if you already trained a model in the previous section and do not want to tune the hyperparameters."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "Hy_aCff_2AML"
},
"outputs": [],
"source": [
"from google.cloud import aiplatform\n",
"from google.cloud.aiplatform import hyperparameter_tuning as hpt\n",
"\n",
"# Init common setup.\n",
"aiplatform.init(project=PROJECT_ID, location=REGION, staging_bucket=GCS_BUCKET)\n",
"\n",
"# Input and output path.\n",
"data_dir = f\"/gcs/{GCS_DATA_DIR}\"\n",
"output_dir = f\"/gcs/{GCS_BUCKET}/timm\"\n",
"\n",
"# Model specific config.\n",
"job_name = f\"pytorch-hp-{MODEL_NAME}\"\n",
"batch_size = 32\n",
"epochs = 2\n",
"\n",
"# Worker pool spec.\n",
"machine_type = \"n1-highmem-16\"\n",
"num_nodes = 1\n",
"gpu_type = \"NVIDIA_TESLA_V100\" # @param {type:\"string\"}\n",
"num_gpus = 2 # @param {type:\"integer\"}\n",
"worker_pool_specs = [\n",
" {\n",
" \"machine_spec\": {\n",
" \"machine_type\": machine_type,\n",
" \"accelerator_type\": gpu_type,\n",
" \"accelerator_count\": num_gpus,\n",
" },\n",
" \"replica_count\": num_nodes,\n",
" \"container_spec\": {\n",
" \"image_uri\": TRAIN_DOCKER_URI,\n",
" \"args\": [\n",
" \"--standalone\",\n",
" f\"--nnodes={num_nodes}\",\n",
" f\"--nproc_per_node={num_gpus}\",\n",
" \"train.py\",\n",
" data_dir,\n",
" f\"--model={MODEL_NAME}\",\n",
" \"--pretrained\",\n",
" f\"--output={output_dir}\",\n",
" f\"--batch-size={batch_size}\",\n",
" f\"--epochs={epochs}\",\n",
" ],\n",
" },\n",
" }\n",
"]\n",
"\n",
"# Hyperparameter job specs.\n",
"metric_spec = {\"top1_accuracy\": \"maximize\"}\n",
"parameter_spec = {\n",
" \"lr\": hpt.DoubleParameterSpec(min=0.001, max=0.05, scale=\"log\"),\n",
"}\n",
"max_trial_count = 2\n",
"parallel_trial_count = 2\n",
"\n",
"# Launch jobs.\n",
"training_job = aiplatform.CustomJob(\n",
" display_name=job_name, worker_pool_specs=worker_pool_specs\n",
")\n",
"hp_job = aiplatform.HyperparameterTuningJob(\n",
" display_name=job_name,\n",
" custom_job=training_job,\n",
" metric_spec=metric_spec,\n",
" parameter_spec=parameter_spec,\n",
" max_trial_count=max_trial_count,\n",
" parallel_trial_count=parallel_trial_count,\n",
")\n",
"hp_job.run()"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "DAyRWwqW2AML"
},
"source": [
"## Deploy model for online prediction\n",
"\n",
"This section uploads the model to Model Registry and deploys it on an Endpoint resource.\n",
"\n",
"The model deployment step will take ~15 minutes to complete."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "jbNbg0yR2AML"
},
"outputs": [],
"source": [
"# The prebuilt serving docker uri.\n",
"SERVE_DOCKER_URI = \"us-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/pytorch-timm-serve\"\n",
"# The port number used by torchserve traffic.\n",
"SERVE_PORT = 7080\n",
"# The path to model checkpoint file, including gs:// prefix.\n",
"MODEL_PT_PATH = \"gs://path_to_model_best.pth.tar\" # @param {type:\"string\"}\n",
"# [Optional] the path to index_to_name.json, including gs:// prefix.\n",
"INDEX_TO_NAME_FILE = \"gs://path_to_index_to_name.json\" # @param {type:\"string\"}"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "INPri3HQ2AML"
},
"source": [
"### Upload and deploy model on Vertex AI"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "WOvqoAaN2AML"
},
"outputs": [],
"source": [
"# Upload model.\n",
"serving_env = {\n",
" \"MODEL_NAME\": MODEL_NAME,\n",
" \"MODEL_PT_PATH\": MODEL_PT_PATH,\n",
" \"INDEX_TO_NAME_FILE\": INDEX_TO_NAME_FILE,\n",
"}\n",
"model = aiplatform.Model.upload(\n",
" display_name=MODEL_NAME,\n",
" serving_container_image_uri=SERVE_DOCKER_URI,\n",
" serving_container_ports=[SERVE_PORT],\n",
" serving_container_predict_route=\"/predictions/timm_serving\",\n",
" serving_container_health_route=\"/ping\",\n",
" serving_container_environment_variables=serving_env,\n",
")\n",
"# Or reuse a pre-uploaded model.\n",
"# model = aiplatform.Model('projects/123456789/locations/us-central1/models/123456789@1')\n",
"\n",
"# Create an endpoint.\n",
"endpoint = aiplatform.Endpoint.create(display_name=\"pytorch-timm-endpoint\")\n",
"# Or reuse a pre-created endpoint.\n",
"# endpoint = aiplatform.Endpoint('projects/123456789/locations/us-central1/endpoints/123456789')\n",
"\n",
"# Deploy model to endpoint.\n",
"model.deploy(\n",
" endpoint=endpoint,\n",
" machine_type=\"n1-standard-8\",\n",
" accelerator_type=\"NVIDIA_TESLA_T4\",\n",
" accelerator_count=1,\n",
" traffic_percentage=100,\n",
" service_account=SERVICE_ACCOUNT,\n",
")"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "s2gnYFEJ2AML"
},
"source": [
"You can mange your uploaded models in the [Model Registry](https://pantheon.corp.google.com/vertex-ai/models) and your endpoints in the [Endpoints](https://pantheon.corp.google.com/vertex-ai/endpoints)."
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "5UP84Q5R2AMM"
},
"source": [
"### Test online prediction\n",
"\n",
"You will now test the deployed endpoint. Please prepare an image to predict."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "zL2Qbm2x2AMM"
},
"outputs": [],
"source": [
"import base64\n",
"\n",
"# You can get the deployed endpoint object by its resource name returned by Endpoint.create(). For example:\n",
"# endpoint = aiplatform.Endpoint('projects/816369962409/locations/us-central1/endpoints/8809168414485512192')\n",
"\n",
"# Please upload an image and enter its filename below.\n",
"IMAGE_FILENAME = \"test.jpg\" # @param {type:\"string\"}\n",
"\n",
"# Alternatively, uncomment the following line to download a cat image for demonstration.\n",
"# ! wget http://images.cocodataset.org/val2017/000000039769.jpg -O test.jpg\n",
"\n",
"with open(IMAGE_FILENAME, \"rb\") as f:\n",
" image_b64 = base64.b64encode(f.read()).decode(\"utf-8\")\n",
"instances = [{\"data\": {\"b64\": image_b64}}]\n",
"\n",
"prediction = endpoint.predict(instances=instances)\n",
"print(prediction)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "KqSxSyT42AMM"
},
"source": [
"### Clean Up Resources"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "tMwdxiH-2AMM"
},
"outputs": [],
"source": [
"endpoint.undeploy_all()\n",
"model.delete()"
]
}
],
"metadata": {
"colab": {
"name": "model_garden_pytorch_timm.ipynb",
"toc_visible": true
},
"kernelspec": {
"display_name": "Python 3",
"name": "python3"
}
},
"nbformat": 4,
"nbformat_minor": 0
}
@@ -332,15 +332,6 @@
")"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "80b3fd2ace09"
},
"source": [
"NOTE: The model weights will be downloaded after the deployment succeeds. Thus additional 5 minutes of waiting time is needed **after** the above model deployment step succeeds and before you run the next step below. Otherwise you might see a `ServiceUnavailable: 503 502:Bad Gateway` error when you send requests to the endpoint."
]
},
{
"cell_type": "code",
"execution_count": null,
@@ -332,15 +332,6 @@
")"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "80b3fd2ace09"
},
"source": [
"NOTE: The model weights will be downloaded after the deployment succeeds. Thus additional 5 minutes of waiting time is needed **after** the above model deployment step succeeds and before you run the next step below. Otherwise you might see a `ServiceUnavailable: 503 502:Bad Gateway` error when you send requests to the endpoint."
]
},
{
"cell_type": "code",
"execution_count": null,
@@ -118,7 +118,7 @@
},
"source": [
"### Colab Only\n",
"Run the following commands for Colab and skip this section if you use Workbench."
"Run the following commands for colab and skip this section if you use workbench."
]
},
{
@@ -181,16 +181,7 @@
"# The project and bucket are for experiments below.\n",
"PROJECT_ID = \"\" # @param {type:\"string\"}\n",
"BUCKET_URI = \"\" # @param {type:\"string\"}\n",
"\n",
"# You can choose a region from https://cloud.google.com/about/locations.\n",
"# Only regions prefixed by \"us\", \"asia\", or \"europe\" are supported.\n",
"REGION = \"us-central1\" # @param {type:\"string\"}\n",
"REGION_PREFIX = REGION.split(\"-\")[0]\n",
"assert REGION_PREFIX in (\n",
" \"us\",\n",
" \"europe\",\n",
" \"asia\",\n",
"), f'{REGION} is not supported. It must be prefixed by \"us\", \"asia\", or \"europe\".'\n",
"REGION = \"us-central1\"\n",
"\n",
"! gcloud config set project $PROJECT_ID\n",
"\n",
@@ -240,13 +231,13 @@
"\n",
"# Data converter constants.\n",
"DATA_CONVERTER_JOB_PREFIX = \"data_converter\"\n",
"DATA_CONVERTER_CONTAINER = f\"{REGION_PREFIX}-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/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 = f\"{REGION_PREFIX}-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/tfvision-oss\"\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",
@@ -256,7 +247,7 @@
"\n",
"# Export constants.\n",
"EXPORT_JOB_PREFIX = \"export\"\n",
"EXPORT_CONTAINER_URI = f\"{REGION_PREFIX}-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/tfvision-serving\"\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",
@@ -265,7 +256,9 @@
"# 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 = f\"{REGION_PREFIX}-docker.pkg.dev/vertex-ai-restricted/prediction/tf_opt-gpu.2-11:latest\"\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",
@@ -314,9 +307,10 @@
" 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\": f\"{location}-aiplatform.googleapis.com\"}\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",
@@ -542,7 +536,7 @@
"# input_train_data_path = ''\n",
"# input_validation_data_path = ''\n",
"\n",
"experiment = \"ResNet-50\" # @param [\"ResNet-50\",\"ResNet-RS-50\",\"Efficientnetv2-m\",\"ViT-ti16\",\"ViT-s16\",\"ViT-b16\",\"ViT-l16\"]\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",
@@ -568,7 +562,6 @@
" **{\n",
" \"experiment\": \"resnet_imagenet\",\n",
" \"config_file\": os.path.join(CONFIG_DIR, \"imagenet_resnet50_gpu.yaml\"),\n",
" \"init_checkpoint\": \"https://storage.googleapis.com/tf_model_garden/vision/resnet/resnet-50-i224.tar.gz\",\n",
" },\n",
" ),\n",
" \"ResNet-RS-50\": dict(\n",
@@ -118,7 +118,7 @@
},
"source": [
"### Colab Only\n",
"Run the following commands for Colab and skip this section if you are using Workbench."
"Run the following commands for colab and skip this section if you use workbench."
]
},
{
@@ -181,37 +181,27 @@
"# The project and bucket are for experiments below.\n",
"PROJECT_ID = \"\" # @param {type:\"string\"}\n",
"BUCKET_URI = \"\" # @param {type:\"string\"}\n",
"\n",
"# You can choose a region from https://cloud.google.com/about/locations.\n",
"# Only regions prefixed by \"us\", \"asia\", or \"europe\" are supported.\n",
"REGION = \"us-central1\" # @param {type:\"string\"}\n",
"REGION_PREFIX = REGION.split(\"-\")[0]\n",
"assert REGION_PREFIX in (\n",
" \"us\",\n",
" \"europe\",\n",
" \"asia\",\n",
"), f'{REGION} is not supported. It must be prefixed by \"us\", \"asia\", or \"europe\".'\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",
"! 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",
"! 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",
"! 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",
"! 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/"
"! gsutil cp scaled_yolov4_1280_gpu.yaml $CONFIG_DIR"
]
},
{
@@ -235,13 +225,13 @@
"\n",
"# Data converter constants.\n",
"DATA_CONVERTER_JOB_PREFIX = \"data_converter\"\n",
"DATA_CONVERTER_CONTAINER = f\"{REGION_PREFIX}-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/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 = f\"{REGION_PREFIX}-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/tfvision-oss\"\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",
@@ -261,7 +251,7 @@
"\n",
"# Export constants.\n",
"EXPORT_JOB_PREFIX = \"export\"\n",
"EXPORT_CONTAINER_URI = f\"{REGION_PREFIX}-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/tfvision-serving\"\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",
@@ -270,7 +260,9 @@
"# 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 = f\"{REGION_PREFIX}-docker.pkg.dev/vertex-ai-restricted/prediction/tf_opt-gpu.2-11:latest\"\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",
@@ -321,9 +313,10 @@
" 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\": f\"{location}-aiplatform.googleapis.com\"}\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",
@@ -476,30 +469,7 @@
" font,\n",
" display_str_list=[display_str],\n",
" )\n",
" return image\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"
" return image"
]
},
{
@@ -651,8 +621,8 @@
" \"objective\": OBJECTIVE,\n",
" \"model_dir\": model_dir,\n",
" \"num_classes\": num_classes,\n",
" \"global_batch_size\": 2,\n",
" \"prefetch_buffer_size\": 6,\n",
" \"global_batch_size\": 4,\n",
" \"prefetch_buffer_size\": 12,\n",
" \"train_steps\": 2000,\n",
" \"input_size\": \"1024,1024\",\n",
"}\n",
@@ -691,19 +661,9 @@
" **{\n",
" \"experiment\": \"scaled_yolo\",\n",
" \"config_file\": TRAIN_YOLOV4_CONFIG,\n",
" \"init_checkpoint\": \"https://storage.googleapis.com/tf_model_garden/vision/yolo/scaled-yolov4/scaled-yolov4-l-p6-i1280.tar.gz\",\n",
" \"input_size\": \"1280,1280\",\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",
"params_override = \"runtime.num_gpus=%s\" % TRAIN_NUM_GPU\n",
"eval_params_override = \"runtime.num_gpus=1,runtime.distribution_strategy=mirrored\"\n",
@@ -721,7 +681,10 @@
" \"--mode=train\",\n",
" \"--params_override=%s\" % params_override,\n",
" ]\n",
" + [\"--{}={}\".format(k, v) for k, v in experiment_container_args.items()],\n",
" + [\n",
" \"--{}={}\".format(k, v)\n",
" for k, v in experiment_container_args_dict[experiment].items()\n",
" ],\n",
" },\n",
" },\n",
" {},\n",
@@ -739,7 +702,10 @@
" \"--mode=continuous_eval\",\n",
" \"--params_override=%s\" % eval_params_override,\n",
" ]\n",
" + [\"--{}={}\".format(k, v) for k, v in experiment_container_args.items()],\n",
" + [\n",
" \"--{}={}\".format(k, v)\n",
" for k, v in experiment_container_args_dict[experiment].items()\n",
" ],\n",
" },\n",
" },\n",
"]\n",
@@ -841,7 +807,8 @@
" \"args\": [\n",
" \"--objective=%s\" % OBJECTIVE,\n",
" \"--input_image_size=1024,1024\",\n",
" \"--experiment=%s\" % experiment_container_args[\"experiment\"],\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",
@@ -119,7 +119,7 @@
"source": [
"### Colab Only\n",
"\n",
"Run the following commands for Colab and skip this section if you use Workbench."
"Run the following commands for colab and skip this section if you use workbench."
]
},
{
@@ -182,28 +182,18 @@
"# The project and bucket are for experiments below.\n",
"PROJECT_ID = \"\" # @param {type:\"string\"}\n",
"BUCKET_URI = \"\" # @param {type:\"string\"}\n",
"\n",
"# You can choose a region from https://cloud.google.com/about/locations.\n",
"# Only regions prefixed by \"us\", \"asia\", or \"europe\" are supported.\n",
"REGION = \"europe-west4\" # @param {type:\"string\"}\n",
"REGION_PREFIX = REGION.split(\"-\")[0]\n",
"assert REGION_PREFIX in (\n",
" \"us\",\n",
" \"europe\",\n",
" \"asia\",\n",
"), f'{REGION} is not supported. It must be prefixed by \"us\", \"asia\", or \"europe\".'\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",
"! 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/"
"! gsutil cp deeplabv3plus_resnet101_cityscapes_gpu_multiworker_mirrored.yaml $CONFIG_DIR"
]
},
{
@@ -227,13 +217,13 @@
"\n",
"# Data converter constants.\n",
"DATA_CONVERTER_JOB_PREFIX = \"data_converter\"\n",
"DATA_CONVERTER_CONTAINER = f\"{REGION_PREFIX}-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/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 = f\"{REGION_PREFIX}-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/tfvision-oss\"\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",
@@ -246,7 +236,7 @@
"\n",
"# Export constants.\n",
"EXPORT_JOB_PREFIX = \"export\"\n",
"EXPORT_CONTAINER_URI = f\"{REGION_PREFIX}-docker.pkg.dev/vertex-ai-restricted/vertex-vision-model-garden-dockers/tfvision-serving\"\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",
@@ -255,7 +245,9 @@
"# 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 = f\"{REGION_PREFIX}-docker.pkg.dev/vertex-ai-restricted/prediction/tf_opt-gpu.2-11:latest\"\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",
@@ -306,9 +298,10 @@
" 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\": f\"{location}-aiplatform.googleapis.com\"}\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",
@@ -515,30 +508,7 @@
" ]\n",
" category_image_color = Image.fromarray(category_image_color_np)\n",
"\n",
" return score_image_grayscale, category_image_color\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"
" return score_image_grayscale, category_image_color"
]
},
{
@@ -695,17 +665,8 @@
" \"prefetch_buffer_size\": 12,\n",
" \"train_steps\": 500,\n",
" \"output_size\": \"1024,2048\",\n",
" \"init_checkpoint\": \"https://storage.googleapis.com/tf_model_garden/vision/deeplabv3plus/dilated-resnet-101-deeplabv3plus.tar.gz\",\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",
@@ -720,7 +681,10 @@
" \"args\": [\n",
" \"--mode=train_and_eval\",\n",
" ]\n",
" + [\"--{}={}\".format(k, v) for k, v in experiment_container_args.items()],\n",
" + [\n",
" \"--{}={}\".format(k, v)\n",
" for k, v in experiment_container_args_dict[experiment].items()\n",
" ],\n",
" },\n",
" },\n",
"]\n",
@@ -823,11 +787,13 @@
" \"command\": [],\n",
" \"args\": [\n",
" \"--objective=%s\" % OBJECTIVE,\n",
" \"--experiment=%s\" % experiment_container_args[\"experiment\"],\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\" % experiment_container_args[\"output_size\"],\n",
" \"--input_image_size=%s\"\n",
" % experiment_container_args_dict[experiment][\"output_size\"],\n",
" ],\n",
" },\n",
" }\n",
@@ -1479,7 +1479,7 @@
"source": [
"For more information about VPC peering in Vertex AI, see https://cloud.google.com/vertex-ai/docs/general/vpc-peering.\n",
"\n",
"**IMPORTANT: you can only setup one VPC peering to servicenetworking.googleapis.com per VPC network.**"
"**IMPORTANT: you can only setup one VPC peering to servicenetworking.googleapis.com per project.**"
]
},
{
+3 -104
View File
@@ -57,77 +57,6 @@ The steps performed are:
&nbsp;&nbsp;&nbsp;Learn more about [BQML ARIMA+ forecasting for tabular data](https://cloud.google.com/vertex-ai/docs/tabular-data/forecasting-arima/overview).
[AutoML training image classification model for batch prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_image_classification_batch_prediction.ipynb)
```
In this tutorial, you create an AutoML image classification model from a Python script, and then do a batch prediction using the Vertex SDK.
The steps performed include:
- Create a Vertex `Dataset` resource.
- Train the model.
- View the model evaluation.
- Make a batch prediction.
```
&nbsp;&nbsp;&nbsp;Learn more about [Get predictions from an image classification model](https://cloud.google.com/vertex-ai/docs/image-data/classification/get-predictions).
[AutoML training image classification model for online prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_image_classification_online_prediction.ipynb)
```
In this tutorial, you create an AutoML image classification 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`.
```
&nbsp;&nbsp;&nbsp;Learn more about [Get predictions from an image classification model](https://cloud.google.com/vertex-ai/docs/image-data/classification/get-predictions).
[AutoML training image object detection model for export to edge](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_image_object_detection_export_edge.ipynb)
```
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.
The steps performed include:
- Create a Vertex `Dataset` resource.
- Train the model.
- Export the `Edge` model from the `Model` resource to Cloud Storage.
- Download the model locally.
- Make a local prediction.
```
[AutoML training image object detection model for online prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_image_object_detection_online_prediction.ipynb)
```
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.
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`.
```
&nbsp;&nbsp;&nbsp;Learn more about [Object detection for image data](https://cloud.google.com/vertex-ai/docs/training-overview#object_detection_for_images).
[AutoML Tabular Workflow pipelines](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_tabular_on_vertex_pipelines.ipynb)
```
@@ -143,38 +72,6 @@ The steps performed are:
&nbsp;&nbsp;&nbsp;Learn more about [Tabular Workflow for E2E AutoML](https://cloud.google.com/vertex-ai/docs/tabular-data/tabular-workflows/e2e-automl).
[AutoML training text entity extraction model for batch prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_text_entity_extraction_batch_prediction.ipynb)
```
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.
The steps performed include:
- Create a Vertex `Dataset` resource.
- Train the model.
- View the model evaluation.
- Make a batch prediction.
```
&nbsp;&nbsp;&nbsp;Learn more about [Entity extraction for text data](https://cloud.google.com/vertex-ai/docs/training-overview#entity_extraction_for_text).
[AutoML training text sentiment analysis model for batch prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/automl_text_sentiment_analysis_batch_prediction.ipynb)
```
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.
The steps performed include:
- Create a Vertex `Dataset` resource.
- Train the model.
- View the model evaluation.
- Make a batch prediction.
```
[Get started with AutoML Training](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/automl/get_started_automl_training.ipynb)
```
@@ -204,7 +101,9 @@ The steps performed include:
- Create a Vertex AI `TimeSeriesDataset` resource.
- Train the model.
- View the model evaluation.
- Make a batch prediction.
- Deploy the `Model` resource to a serving `Endpoint` resource.
- Make a prediction.
- Undeploy the `Model`.
```
@@ -123,6 +123,39 @@
"to generate a cost estimate based on your projected usage."
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "setup_local"
},
"source": [
"### Set up your local development environment\n",
"\n",
"If you are using Colab or Google Cloud Notebooks, your environment already meets all the requirements to run this notebook. You can skip this step.\n",
"\n",
"Otherwise, make sure your environment meets this notebook's requirements. You need the following:\n",
"\n",
"- The Cloud Storage SDK\n",
"- Git\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. To install Jupyter, run `pip3 install jupyter` on the command-line in a terminal shell.\n",
"\n",
"5. To launch Jupyter, run `jupyter notebook` on the command-line in a terminal shell.\n",
"\n",
"6. Open this notebook in the Jupyter Notebook Dashboard.\n"
]
},
{
"cell_type": "markdown",
"metadata": {
@@ -131,7 +164,7 @@
"source": [
"## Installation\n",
"\n",
"Install the latest version of Vertex AI SDK for Python."
"Install the latest version of Vertex SDK for Python."
]
},
{
@@ -144,8 +177,33 @@
"source": [
"import os\n",
"\n",
"! pip3 install --upgrade --quiet google-cloud-aiplatform \\\n",
" google-cloud-storage"
"# 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"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "install_storage"
},
"source": [
"Install the latest GA version of *google-cloud-storage* library as well."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "install_storage"
},
"outputs": [],
"source": [
"! pip3 install -U google-cloud-storage $USER_FLAG"
]
},
{
@@ -154,38 +212,57 @@
"id": "restart"
},
"source": [
"### Colab only: Uncomment the following cell to restart the kernel"
"### Restart the kernel\n",
"\n",
"Once you've installed the additional packages, you need to restart the notebook kernel so it can find the packages."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "D-ZBOjErv5mM"
"id": "restart"
},
"outputs": [],
"source": [
"# Automatically restart kernel after installs so that your environment can access the new packages\n",
"# import IPython\n",
"import os\n",
"\n",
"# app = IPython.Application.instance()\n",
"# app.kernel.do_shutdown(True)"
"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": "yfEglUHQk9S3"
"id": "before_you_begin:nogpu"
},
"source": [
"## Before you begin\n",
"\n",
"### Set your project ID\n",
"### GPU runtime\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)"
"This tutorial does not require a GPU runtime.\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 $300 free credit towards your compute/storage costs.\n",
"\n",
"2. [Make sure that billing is enabled for your project.](https://cloud.google.com/billing/docs/how-to/modify-project)\n",
"\n",
"3. [Enable the following APIs: Vertex AI APIs, Compute Engine APIs, and Cloud Storage.](https://console.cloud.google.com/flows/enableapi?apiid=ml.googleapis.com,compute_component,storage-component.googleapis.com)\n",
"\n",
"4. If you are running this notebook locally, you will need to install the [Cloud SDK]((https://cloud.google.com/sdk)).\n",
"\n",
"5. Enter your project ID in the cell below. Then run the cell to make sure the\n",
"Cloud SDK uses the right project for all the commands in this notebook.\n",
"\n",
"**Note**: Jupyter runs lines prefixed with `!` as shell commands, and it interpolates Python variables prefixed with `$`."
]
},
{
@@ -196,10 +273,33 @@
},
"outputs": [],
"source": [
"PROJECT_ID = \"[your-project-id]\" # @param {type:\"string\"}\n",
"\n",
"# Set the project id\n",
"! gcloud config set project {PROJECT_ID}"
"PROJECT_ID = \"[your-project-id]\" # @param {type:\"string\"}"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "autoset_project_id"
},
"outputs": [],
"source": [
"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": "set_gcloud_project_id"
},
"outputs": [],
"source": [
"! gcloud config set project $PROJECT_ID"
]
},
{
@@ -210,7 +310,16 @@
"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)."
"You can also change the `REGION` variable, which is used for operations\n",
"throughout the rest of this notebook. Below are regions supported for Vertex AI. We recommend that you choose the region closest to you.\n",
"\n",
"- Americas: `us-central1`\n",
"- Europe: `europe-west4`\n",
"- Asia Pacific: `asia-east1`\n",
"\n",
"You may not use a multi-regional bucket for training with Vertex AI. Not all regions provide support for all Vertex AI services.\n",
"\n",
"Learn more about [Vertex AI regions](https://cloud.google.com/vertex-ai/docs/general/locations)"
]
},
{
@@ -221,7 +330,33 @@
},
"outputs": [],
"source": [
"REGION = \"us-central1\" # @param {type: \"string\"}"
"REGION = \"[your-region]\" # @param {type: \"string\"}\n",
"if REGION == \"[your-region]\":\n",
" REGION = \"us-central1\""
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "timestamp"
},
"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 the timestamp onto the name of resources you create in this tutorial."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "timestamp"
},
"outputs": [],
"source": [
"from datetime import datetime\n",
"\n",
"TIMESTAMP = datetime.now().strftime(\"%Y%m%d%H%M%S\")"
]
},
{
@@ -232,54 +367,53 @@
"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",
"**If you are using Google Cloud Notebooks**, your environment is already authenticated. Skip this step.\n",
"\n",
"**1. Vertex AI Workbench**\n",
"* Do nothing as you are already authenticated.\n",
"**If you are using Colab**, run the cell below and follow the instructions when prompted to authenticate your account via oAuth.\n",
"\n",
"**2. Local JupyterLab instance, uncomment and run:**"
"**Otherwise**, follow these steps:\n",
"\n",
"In the Cloud Console, go to the [Create service account key](https://console.cloud.google.com/apis/credentials/serviceaccountkey) page.\n",
"\n",
"**Click Create service account**.\n",
"\n",
"In the **Service account name** field, enter a name, and click **Create**.\n",
"\n",
"In the **Grant this service account access to project** section, click the Role drop-down list. Type \"Vertex\" into the filter box, and select **Vertex Administrator**. Type \"Storage Object Admin\" into the filter box, and select **Storage Object Admin**.\n",
"\n",
"Click Create. A JSON file that contains your key downloads to your local environment.\n",
"\n",
"Enter the path to your service account key as the GOOGLE_APPLICATION_CREDENTIALS variable in the cell below and run the cell."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "ce6043da7b33"
"id": "gcp_authenticate"
},
"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."
"# 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 Google Cloud Notebook, then don't execute this code\n",
"if not os.path.exists(\"/opt/deeplearning/metadata/env_version\"):\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 ''"
]
},
{
@@ -290,7 +424,11 @@
"source": [
"### Create a Cloud Storage bucket\n",
"\n",
"Create a storage bucket to store intermediate artifacts such as datasets."
"**The following steps are required, regardless of your notebook environment.**\n",
"\n",
"When you initialize the Vertex 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."
]
},
{
@@ -301,7 +439,19 @@
},
"outputs": [],
"source": [
"BUCKET_URI = f\"gs://your-bucket-name-{PROJECT_ID}-unique\" # @param {type:\"string\"}"
"BUCKET_NAME = \"gs://[your-bucket-name]\" # @param {type:\"string\"}"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "autoset_bucket"
},
"outputs": [],
"source": [
"if BUCKET_NAME == \"\" or BUCKET_NAME is None or BUCKET_NAME == \"gs://[your-bucket-name]\":\n",
" BUCKET_NAME = \"gs://\" + PROJECT_ID + \"aip-\" + TIMESTAMP"
]
},
{
@@ -321,7 +471,27 @@
},
"outputs": [],
"source": [
"! gsutil mb -l $REGION -p $PROJECT_ID $BUCKET_URI"
"! gsutil mb -l $REGION -p $PROJECT_ID $BUCKET_NAME"
]
},
{
"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": "validate_bucket"
},
"outputs": [],
"source": [
"! gsutil ls -al $BUCKET_NAME"
]
},
{
@@ -368,7 +538,7 @@
},
"outputs": [],
"source": [
"aiplatform.init(project=PROJECT_ID, staging_bucket=BUCKET_URI)"
"aiplatform.init(project=PROJECT_ID, staging_bucket=BUCKET_NAME)"
]
},
{
@@ -460,7 +630,7 @@
"outputs": [],
"source": [
"dataset = aiplatform.TextDataset.create(\n",
" display_name=\"NCBI Biomedical\",\n",
" display_name=\"NCBI Biomedical\" + \"_\" + TIMESTAMP,\n",
" gcs_source=[IMPORT_FILE],\n",
" import_schema_uri=aiplatform.schema.dataset.ioformat.text.extraction,\n",
")\n",
@@ -500,7 +670,7 @@
"outputs": [],
"source": [
"job = aiplatform.AutoMLTextTrainingJob(\n",
" display_name=\"biomedical\", prediction_type=\"extraction\"\n",
" display_name=\"biomedical_\" + TIMESTAMP, prediction_type=\"extraction\"\n",
")\n",
"\n",
"print(job)"
@@ -537,7 +707,7 @@
"source": [
"model = job.run(\n",
" dataset=dataset,\n",
" model_display_name=\"biomedical\",\n",
" model_display_name=\"biomedical_\" + TIMESTAMP,\n",
" training_fraction_split=0.8,\n",
" validation_fraction_split=0.1,\n",
" test_fraction_split=0.1,\n",
@@ -565,7 +735,7 @@
"outputs": [],
"source": [
"# Get model resource ID\n",
"models = aiplatform.Model.list(filter=\"display_name=biomedical\")\n",
"models = aiplatform.Model.list(filter=\"display_name=biomedical_\" + TIMESTAMP)\n",
"\n",
"# Get a reference to the Model Service client\n",
"client_options = {\"api_endpoint\": f\"{REGION}-aiplatform.googleapis.com\"}\n",
@@ -746,7 +916,7 @@
"job.delete()\n",
"\n",
"if os.getenv(\"IS_TESTING\"):\n",
" ! gsutil -m rm -r $BUCKET_URI"
" ! gsutil -m rm -r $BUCKET_NAME"
]
}
],
-55
View File
@@ -60,22 +60,6 @@ The steps performed include:
&nbsp;&nbsp;&nbsp;Learn more about [Vertex AI Training](https://cloud.google.com/vertex-ai/docs/training/custom-training).
[Custom training with custom training container and automatic registering of the model](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/custom/custom_training_container_and_model_registry.ipynb)
```
In this tutorial, you create a custom model from a Python script in a custom Docker container using the Vertex AI SDK, and automatically register the model in the Vertex AI Model Registry.
The steps performed include:
- Create a Vertex AI custom job for training a model.
- Train and register a TensorFlow model using a custom container,
- List the registered model from the Vertex AI Model Registry.
```
&nbsp;&nbsp;&nbsp;Learn more about [Custom 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)
```
@@ -93,45 +77,6 @@ The steps performed include:
&nbsp;&nbsp;&nbsp;Learn more about [Vertex AI TensorBoard Profiler](https://cloud.google.com/vertex-ai/docs/experiments/tensorboard-profiler).
[Get started with Vertex AI Training for XGBoost](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/custom/get_started_vertex_training.ipynb)
```
Learn how to use `Vertex AI Training` for training a XGBoost custom model.
The steps performed include:
- Training using a Python package.
- Report accuracy when hyperparameter tuning.
- Save the model artifacts to Cloud Storage using GCSFuse.
- Create a `Vertex AI Model` resource.
```
&nbsp;&nbsp;&nbsp;Learn more about [Custom training](https://cloud.google.com/vertex-ai/docs/training/custom-training).
[Get started with Endpoint and shared VM](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/custom/get_started_with_vertex_endpoint_and_shared_vm.ipynb)
```
Learn how to use deployment resource pools for deploying models.
The steps performed include:
- Upload a pre-trained image classification model as a `Model` resource (model A).
- Upload a pre-trained text sentence encoder model as a `Model` resource (model B).
- Create a shared VM deployment resource pool.
- List shared VM deployment resource pools.
- Create two `Endpoint` resources.
- Deploy first model (model A) to first `Endpoint` resource using deployment resource pool.
- Deploy second model (model B) to second `Endpoint` resource using deployment resource pool.
- Make a prediction request with first deployed model (model A).
- Make a prediction request with second deployed model (model B).
```
&nbsp;&nbsp;&nbsp;Learn more about [Shared resources across deployments](https://cloud.google.com/vertex-ai/docs/predictions/model-co-hosting).
[Custom training and batch prediction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/custom/sdk-custom-image-classification-batch.ipynb)
```
-2
View File
@@ -18,8 +18,6 @@ The steps performed include:
&nbsp;&nbsp;&nbsp;Learn more about [BigQuery Datasets](https://cloud.google.com/bigquery/docs/datasets-intro).
&nbsp;&nbsp;&nbsp;Learn more about [Vertex AI for BigQuery users](https://cloud.google.com/vertex-ai/docs/beginner/bqml).
[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)
-33
View File
@@ -56,24 +56,6 @@ The steps performed include:
&nbsp;&nbsp;&nbsp;Learn more about [Vertex AI Pipelines](https://cloud.google.com/vertex-ai/docs/pipelines/introduction).
[Delete Outdated Experiments in Vertex AI TensorBoard](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/experiments/delete_outdated_tensorboard_experiments.ipynb)
```
Learn how to delete outdated TensorBoard Experiments to avoid unnecessary storage costs.
The steps performed include:
- How to delete the TB Experiment with a predefined key-value label pair `<label_key, label_value>`
- How to delete the TB Experiments created before the `create_time`
- How to delete the TB Experiments created before the `update_time`
```
&nbsp;&nbsp;&nbsp;Learn more about [Vertex AI TensorBoard](https://cloud.google.com/vertex-ai/docs/experiments/tensorboard-overview).
[Get started with Vertex AI Experiments](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/experiments/get_started_with_vertex_experiments.ipynb)
```
@@ -106,18 +88,3 @@ The steps performed include:
&nbsp;&nbsp;&nbsp;Learn more about [Custom training](https://cloud.google.com/vertex-ai/docs/training/custom-training).
[Autologging](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/experiments/autologging.ipynb)
```
Learn how to use `Vertex AI Autologging`.
The steps performed include:
- Enable autologging in the Vertex AI SDK.
- Train scikit-learn model and see the resulting experiment run with metrics and parameters autologged to Vertex AI Experiments without setting an experiment run.
- Train Tensorflow model, check autologged metrics and parameters to Vertex AI Experiments by manually setting an experiment run with `aiplatform.start_run()` and `aiplatform.end_run()`.
- Disable autologging in the Vertex AI SDK, train a PyTorch model and check that none of the parameters or metrics are logged.
```
@@ -628,7 +628,6 @@
"outputs": [],
"source": [
"import matplotlib.pyplot as plt\n",
"import numpy as np\n",
"import pandas as pd\n",
"from google.cloud import aiplatform as vertex_ai\n",
"from tensorflow.python.keras import Sequential, layers\n",
@@ -968,12 +967,6 @@
"\n",
" # Log final metrics\n",
" loss, mae, mse = model.evaluate(normed_test_data, test_labels, verbose=2)\n",
" if np.isnan(loss):\n",
" loss = 0\n",
" if np.isnan(mae):\n",
" mae = 0\n",
" if np.isnan(mse):\n",
" mse = 0\n",
" vertex_ai.log_metrics({\"eval_loss\": loss, \"eval_mae\": mae, \"eval_mse\": mse})\n",
"\n",
" vertex_ai.end_run()"
@@ -192,7 +192,7 @@
"if IS_WORKBENCH_NOTEBOOK:\n",
" USER_FLAG = \"--user\"\n",
"\n",
"! pip3 install --upgrade google-cloud-aiplatform pandas {USER_FLAG} -q"
"! pip3 install --upgrade google-cloud-aiplatform {USER_FLAG} -q"
]
},
{
@@ -34,18 +34,18 @@
"<table align=\"left\">\n",
"\n",
" <td>\n",
" <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/experiments/get_started_with_vertex_experiments_autologging.ipynb\">\n",
" <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/experiments/autologging.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/experiments/get_started_with_vertex_experiments_autologging.ipynb\">\n",
" <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/experiments/autologging.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/experiments/get_started_with_vertex_experiments_autologging.ipynb\">\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/experiments/autologging.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",
+1 -1
View File
@@ -60,7 +60,7 @@ The steps performed include:
&nbsp;&nbsp;&nbsp;Learn more about [Vertex AI Batch Prediction](https://cloud.google.com/vertex-ai/docs/tabular-data/classification-regression/get-batch-predictions).
[Custom training image classification model for online prediction with explainability](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/explainable_ai/sdk_custom_image_classification_online_explain.ipynb)
[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.
File diff suppressed because one or more lines are too long
+110 -110
View File
@@ -1,5 +1,5 @@
[AutoML Image Classification](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-automl-image-classification-batch-online.ipynb)
[AutoML Image Classification](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ1 Vertex SDK AutoML Image Classification.ipynb)
```
Learn to use `AutoML` to train an image model and use `Vertex AI Prediction` and `Vertex AI Batch Prediction` to do online and batch predictions.
@@ -18,14 +18,15 @@ The steps performed include:
&nbsp;&nbsp;&nbsp;Learn more about [Classification for image data](https://cloud.google.com/vertex-ai/docs/training-overview#classification_for_images).
[AutoML image object detection](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-automl-image-object-detection-batch-online.ipynb)
[Custom Scikit-Learn model with pre-built training container](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ10 Vertex SDK Custom Scikit-Learn with pre-built training container.ipynb)
```
Learn to use `AutoML` to train an image model and use `Vertex AI Prediction` and `Vertex AI Batch Prediction` to do online and batch predictions.
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:
- Train an AutoML object detection model.
- Create a `Vertex AI` custom job for training a scikit-learn model.
- Upload the trained model artifacts as a `Model` resource.
- Make a batch prediction.
- Deploy model to a endpoint
- Make a online prediction
@@ -34,7 +35,40 @@ The steps performed include:
&nbsp;&nbsp;&nbsp;Learn more about [Migrate to Vertex AI](https://cloud.google.com/vertex-ai/docs/start/migrating-to-vertex-ai).
&nbsp;&nbsp;&nbsp;Learn more about [Object detection for image data](https://cloud.google.com/vertex-ai/docs/training-overview#object_detection_for_images).
&nbsp;&nbsp;&nbsp;Learn more about [Custom training](https://cloud.google.com/vertex-ai/docs/training/custom-training).
[Hyperparameter Tuning](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ11 Vertex SDK Hyperparameter Tuning.ipynb)
```
Learn to use `Vertex AI Hyperparameter` to create and tune a custom trained model.
The steps performed include:
- Create a `Vertex AI` hyperparameter tuning job for training a TensorFlow model.
```
&nbsp;&nbsp;&nbsp;Learn more about [Migrate to Vertex AI](https://cloud.google.com/vertex-ai/docs/training/hyperparameter-tuning-overview).
&nbsp;&nbsp;&nbsp;Learn more about [Custom training](https://cloud.google.com/vertex-ai/docs/training/custom-training).
[AutoML Video Classification](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ14 Vertex SDK AutoML Video Classification.ipynb)
```
Learn to use `AutoML` to train a video model and use `Vertex AI Batch Prediction` to do batch predictions.
The steps performed include:
- Train an AutoML video classification model.
- Make a batch prediction.
```
&nbsp;&nbsp;&nbsp;Learn more about [Migrate to Vertex AI](https://cloud.google.com/vertex-ai/docs/start/migrating-to-vertex-ai).
&nbsp;&nbsp;&nbsp;Learn more about [Classification for video data](https://cloud.google.com/vertex-ai/docs/training-overview#classification_for_videos).
[AutoML Video Object Tracking](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ15 Vertex SDK AutoML Object Tracking.ipynb)
@@ -54,7 +88,55 @@ The steps performed include:
&nbsp;&nbsp;&nbsp;Learn more about [Object tracking for video data](https://cloud.google.com/vertex-ai/docs/training-overview#object_tracking_for_videos).
[AutoML tabular binary classification](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-automl-tabular-binary-classification-online-prediction.ipynb)
[Custom Image Classification w/pre-built training container](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ2,12 Vertex SDK Custom Image Classification with pre-built training container.ipynb)
```
Learn how to train a tensorflow image classification model using a prebuilt container and Vertex AI training.
The steps performed include:
- *Package the training code into a python application.*
- *Containerize the training application using Cloud Build and Artifact Registry.*
- *Create a custom container training job in Vertex AI and run it.*
- *Evaluate the model generated from the training job.*
- *Create a model resource for the trained model in Vertex AI Model Registry.*
- *Run a Vertex AI batch prediction job.*
- *Deploy the model resource to a Vertex AI Endpoint.*
- *Run a online prediction job on the model resource.*
- *Clean up the resources created.*
```
&nbsp;&nbsp;&nbsp;Learn more about [Migrate to Vertex AI](https://cloud.google.com/vertex-ai/docs/start/migrating-to-vertex-ai).
&nbsp;&nbsp;&nbsp;Learn more about [Custom training](https://cloud.google.com/vertex-ai/docs/training/custom-training).
[Custom Image Classification w/custom training container](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ3 Vertex SDK Custom Image Classification with custom training container.ipynb)
```
Learn how to train a tensorflow image classification model using a custom container and Vertex AI training.
The steps performed include:
- *Package the training code into a python application.*
- *Containerize the training application using Cloud Build and Artifact Registry.*
- *Create a custom container training job in Vertex AI and run it.*
- *Evaluate the model generated from the training job.*
- *Create a model resource for the trained model in Vertex AI Model Registry.*
- *Run a Vertex AI batch prediction job.*
- *Deploy the model resource to a Vertex AI Endpoint.*
- *Run a online prediction job on the model resource.*
- *Clean up the resources created.*
```
&nbsp;&nbsp;&nbsp;Learn more about [Migrate to Vertex AI](https://cloud.google.com/vertex-ai/docs/start/migrating-to-vertex-ai).
&nbsp;&nbsp;&nbsp;Learn more about [Custom training](https://cloud.google.com/vertex-ai/docs/training/custom-training).
[AutoML Tabular Binary Classification](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ4 Vertex SDK AutoML Tabular Binary Classification.ipynb)
```
In this tutorial, you create an AutoML tabular binary classification model and deploy for online prediction from a Python script using the Vertex AI SDK.
@@ -75,7 +157,26 @@ The steps performed include:
&nbsp;&nbsp;&nbsp;Learn more about [Classification for tabular data](https://cloud.google.com/vertex-ai/docs/tabular-data/classification-regression/overview).
[AutoML Text Classification](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-automl-text-classification-batch-prediction.ipynb)
[AutoML Image Object Detection](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ5 Vertex SDK AutoML Image Object Detection.ipynb)
```
Learn to use `AutoML` to train an image model and use `Vertex AI Prediction` and `Vertex AI Batch Prediction` to do online and batch predictions.
The steps performed include:
- Train an AutoML object detection model.
- Make a batch prediction.
- Deploy model to a endpoint
- Make a online prediction
```
&nbsp;&nbsp;&nbsp;Learn more about [Migrate to Vertex AI](https://cloud.google.com/vertex-ai/docs/start/migrating-to-vertex-ai).
&nbsp;&nbsp;&nbsp;Learn more about [Object detection for image data](https://cloud.google.com/vertex-ai/docs/training-overview#object_detection_for_images).
[AutoML Text Classification](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ6 Vertex SDK AutoML Text Classification.ipynb)
```
The objective of this notebook is to build a AutoML Text Classification Model.
@@ -97,7 +198,7 @@ The steps performed include the following:
&nbsp;&nbsp;&nbsp;Learn more about [Classification for text data](https://cloud.google.com/vertex-ai/docs/training-overview#classification_for_text).
[AutoML Text Entity Extraction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-automl-text-classification-batch-prediction.ipynb)
[AutoML Text Entity Extraction](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ7 Vertex SDK AutoML Text Entity Extraction.ipynb)
```
The objective of this notebook is to build a AutoML Text Entity Extraction model.
@@ -140,72 +241,7 @@ The steps performed include the following:
&nbsp;&nbsp;&nbsp;Learn more about [Sentiment analysis for text data](https://cloud.google.com/vertex-ai/docs/training-overview#sentiment_analysis_for_text).
[AutoML Video Classification](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ14 Vertex SDK AutoML Video Classification.ipynb)
```
Learn to use `AutoML` to train a video model and use `Vertex AI Batch Prediction` to do batch predictions.
The steps performed include:
- Train an AutoML video classification model.
- Make a batch prediction.
```
&nbsp;&nbsp;&nbsp;Learn more about [Migrate to Vertex AI](https://cloud.google.com/vertex-ai/docs/start/migrating-to-vertex-ai).
&nbsp;&nbsp;&nbsp;Learn more about [Classification for video data](https://cloud.google.com/vertex-ai/docs/training-overview#classification_for_videos).
[Custom image classification with a custom training container](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-custom-image-classification-custom-container.ipynb)
```
Learn how to train a tensorflow image classification model using a custom container and Vertex AI training.
The steps performed include:
- *Package the training code into a python application.*
- *Containerize the training application using Cloud Build and Artifact Registry.*
- *Create a custom container training job in Vertex AI and run it.*
- *Evaluate the model generated from the training job.*
- *Create a model resource for the trained model in Vertex AI Model Registry.*
- *Run a Vertex AI batch prediction job.*
- *Deploy the model resource to a Vertex AI Endpoint.*
- *Run a online prediction job on the model resource.*
- *Clean up the resources created.*
```
&nbsp;&nbsp;&nbsp;Learn more about [Migrate to Vertex AI](https://cloud.google.com/vertex-ai/docs/start/migrating-to-vertex-ai).
&nbsp;&nbsp;&nbsp;Learn more about [Custom training](https://cloud.google.com/vertex-ai/docs/training/custom-training).
[Custom image classification with a pre-built training container](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-custom-image-classification-prebuilt-container.ipynb)
```
Learn how to train a tensorflow image classification model using a prebuilt container and Vertex AI training.
The steps performed include:
- *Package the training code into a python application.*
- *Containerize the training application using Cloud Build and Artifact Registry.*
- *Create a custom container training job in Vertex AI and run it.*
- *Evaluate the model generated from the training job.*
- *Create a model resource for the trained model in Vertex AI Model Registry.*
- *Run a Vertex AI batch prediction job.*
- *Deploy the model resource to a Vertex AI Endpoint.*
- *Run a online prediction job on the model resource.*
- *Clean up the resources created.*
```
&nbsp;&nbsp;&nbsp;Learn more about [Migrate to Vertex AI](https://cloud.google.com/vertex-ai/docs/start/migrating-to-vertex-ai).
&nbsp;&nbsp;&nbsp;Learn more about [Custom training](https://cloud.google.com/vertex-ai/docs/training/custom-training).
[Custom Scikit-Learn model with pre-built training container](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-custom-scikit-learn-prebuilt-container.ipynb)
[Custom XGBoost model with pre-built training container](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ9 Vertex SDK Custom XGBoost with pre-built training container.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.
@@ -224,39 +260,3 @@ The steps performed include:
&nbsp;&nbsp;&nbsp;Learn more about [Custom training](https://cloud.google.com/vertex-ai/docs/training/custom-training).
[Custom XGBoost model with pre-built training container](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-custom-xgboost-prebuilt-container.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 scikit-learn model.
- Upload the trained model artifacts as a `Model` resource.
- Make a batch prediction.
- Deploy model to a endpoint
- Make a online prediction
```
&nbsp;&nbsp;&nbsp;Learn more about [Migrate to Vertex AI](https://cloud.google.com/vertex-ai/docs/start/migrating-to-vertex-ai).
&nbsp;&nbsp;&nbsp;Learn more about [Custom training](https://cloud.google.com/vertex-ai/docs/training/custom-training).
[Hyperparameter Tuning](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-hyperparameter-tuning.ipynb)
```
Learn to use `Vertex AI Hyperparameter` to create and tune a custom trained model.
The steps performed include:
- Create a `Vertex AI` hyperparameter tuning job for training a TensorFlow model.
```
&nbsp;&nbsp;&nbsp;Learn more about [Migrate to Vertex AI](https://cloud.google.com/vertex-ai/docs/training/hyperparameter-tuning-overview).
&nbsp;&nbsp;&nbsp;Learn more about [Custom training](https://cloud.google.com/vertex-ai/docs/training/custom-training).
@@ -34,18 +34,18 @@
"<table align=\"left\">\n",
"\n",
" <td>\n",
" <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-automl-image-classification-batch-online.ipynb\">\n",
" <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ1 Vertex SDK AutoML 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",
" <td>\n",
" <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-automl-image-classification-batch-online.ipynb\">\n",
" <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ1 Vertex SDK AutoML 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>\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/migration/sdk-automl-image-classification-batch-online.ipynb\" target='_blank'>\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/migration/UJ1 Vertex SDK AutoML Image Classification.ipynb\" target='_blank'>\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",
@@ -1458,7 +1458,7 @@
],
"metadata": {
"colab": {
"name": "sdk-automl-image-classification-batch-online.ipynb",
"name": "UJ1 Vertex SDK AutoML Image Classification.ipynb",
"toc_visible": true
},
"kernelspec": {
@@ -34,18 +34,18 @@
"<table align=\"left\">\n",
"\n",
" <td>\n",
" <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-hyperparameter-tuning.ipynb\">\n",
" <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ11 Vertex SDK Hyperparameter Tuning.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/migration/sdk-hyperparameter-tuning.ipynb\">\n",
" <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ11 Vertex SDK Hyperparameter Tuning.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/migration/sdk-hyperparameter-tuning.ipynb\" target='_blank'>\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/migration/UJ11 Vertex SDK Hyperparameter Tuning.ipynb\" target='_blank'>\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",
@@ -1414,7 +1414,7 @@
],
"metadata": {
"colab": {
"name": "sdk-hyperparameter-tuning.ipynb",
"name": "UJ11 Vertex SDK Hyperparameter Tuning.ipynb",
"toc_visible": true
},
"kernelspec": {
@@ -1297,7 +1297,7 @@
],
"metadata": {
"colab": {
"name": "sdk-automl-video-classification-batch-prediction.ipynb",
"name": "UJ14 Vertex SDK AutoML Video Classification.ipynb",
"toc_visible": true
},
"kernelspec": {
@@ -633,9 +633,8 @@
" gcs_source=[IMPORT_FILE],\n",
" import_schema_uri=aip.schema.dataset.ioformat.video.object_tracking,\n",
")\n",
"if os.getenv(\"IS_TESTING\"):\n",
"if os.getenv('IS_TESTING'):\n",
" import time\n",
"\n",
" time.sleep(30)\n",
"\n",
"print(dataset.resource_name)"
@@ -1252,7 +1251,7 @@
],
"metadata": {
"colab": {
"name": "sdk-automl-object-tracking-batch-prediction.ipynb",
"name": "UJ15 Vertex SDK AutoML Object Tracking.ipynb",
"toc_visible": true
},
"kernelspec": {
@@ -34,18 +34,18 @@
"<table align=\"left\">\n",
"\n",
" <td>\n",
" <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-custom-image-classification-prebuilt-container.ipynb\">\n",
" <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ2,12 Vertex SDK Custom Image Classification with pre-built training container.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/migration/sdk-custom-image-classification-prebuilt-container.ipynb\">\n",
" <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ2,12 Vertex SDK Custom Image Classification with pre-built training container.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/migration/sdk-custom-image-classification-prebuilt-container.ipynb\" target='_blank'>\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/migration/UJ2,12 Vertex SDK Custom Image Classification with pre-built training container.ipynb\" target='_blank'>\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",
@@ -2005,7 +2005,7 @@
],
"metadata": {
"colab": {
"name": "sdk-custom-image-classification-prebuilt-container.ipynb",
"name": "UJ2,12 Vertex SDK Custom Image Classification with pre-built training container.ipynb",
"toc_visible": true
},
"kernelspec": {
@@ -34,18 +34,18 @@
"<table align=\"left\">\n",
"\n",
" <td>\n",
"<a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-custom-image-classification-custom-container.ipynb\" target='_blank'>\n",
"<a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ3 Vertex SDK Custom Image Classification with custom training container.ipynb\" target='_blank'>\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/migration/sdk-custom-image-classification-custom-container.ipynb\" target='_blank'>\n",
"<a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ3 Vertex SDK Custom Image Classification with custom training container.ipynb\" target='_blank'>\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/migration/sdk-custom-image-classification-custom-container.ipynb\" target='_blank'>\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/migration/UJ3 Vertex SDK Custom Image Classification with custom training container.ipynb\" target='_blank'>\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",
@@ -1899,7 +1899,7 @@
],
"metadata": {
"colab": {
"name": "sdk-custom-image-classification-custom-container.ipynb",
"name": "UJ3 Vertex SDK Custom Image Classification with custom training container.ipynb",
"toc_visible": true
},
"kernelspec": {
@@ -34,18 +34,18 @@
"<table align=\"left\">\n",
"\n",
" <td>\n",
" <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-automl-tabular-binary-classification-online-prediction.ipynb\">\n",
" <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ4 Vertex SDK AutoML Tabular Binary 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",
" <td>\n",
" <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-automl-tabular-binary-classification-online-prediction.ipynb\">\n",
" <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ4 Vertex SDK AutoML Tabular Binary 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>\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/migration/sdk-automl-tabular-binary-classification-online-prediction.ipynb\">\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/migration/UJ4 Vertex SDK AutoML Tabular Binary 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",
@@ -1410,7 +1410,7 @@
],
"metadata": {
"colab": {
"name": "sdk-automl-tabular-binary-classification-online-prediction.ipynb",
"name": "UJ4 Vertex SDK AutoML Tabular Binary Classification.ipynb",
"toc_visible": true
},
"kernelspec": {
@@ -34,18 +34,18 @@
"<table align=\"left\">\n",
"\n",
" <td>\n",
" <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-automl-image-object-detection-batch-online.ipynb\">\n",
" <a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ5 Vertex SDK AutoML 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",
" <td>\n",
" <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-automl-image-object-detection-batch-online.ipynb\">\n",
" <a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ5 Vertex SDK AutoML 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>\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/migration/sdk-automl-image-object-detection-batch-online.ipynb\" target='_blank'>\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/migration/UJ5 Vertex SDK AutoML Image Object Detection.ipynb\" target='_blank'>\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",
@@ -1454,7 +1454,7 @@
],
"metadata": {
"colab": {
"name": "sdk-automl-image-object-detection-batch-online.ipynb",
"name": "UJ5 Vertex SDK AutoML Image Object Detection.ipynb",
"toc_visible": true
},
"kernelspec": {
@@ -33,18 +33,18 @@
"\n",
"<table align=\"left\">\n",
" <td>\n",
"<a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-automl-text-classification-batch-prediction.ipynb\" target='_blank'>\n",
"<a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ6 Vertex SDK AutoML Text Classification.ipynb\" target='_blank'>\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/migration/sdk-automl-text-classification-batch-prediction.ipynb\" target='_blank'>\n",
"<a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ6 Vertex SDK AutoML Text Classification.ipynb\" target='_blank'>\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/migration/sdk-automl-text-classification-batch-prediction.ipynb\" target='_blank'>\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/migration/UJ6 Vertex SDK AutoML Text Classification.ipynb\" target='_blank'>\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",
@@ -1337,7 +1337,7 @@
],
"metadata": {
"colab": {
"name": "sdk-automl-text-classification-batch-prediction.ipynb",
"name": "UJ6 Vertex SDK AutoML Text Classification.ipynb",
"toc_visible": true
},
"kernelspec": {
@@ -33,18 +33,18 @@
"\n",
"<table align=\"left\">\n",
" <td>\n",
"<a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-automl-text-classification-batch-prediction.ipynb\" target='_blank'>\n",
"<a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ7 Vertex SDK AutoML Text Entity Extraction.ipynb\" target='_blank'>\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/migration/sdk-automl-text-classification-batch-prediction.ipynb\" target='_blank'>\n",
"<a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ7 Vertex SDK AutoML Text Entity Extraction.ipynb\" target='_blank'>\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/migration/sdk-automl-text-classification-batch-prediction.ipynb\" target='_blank'>\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/migration/UJ7 Vertex SDK AutoML Text Entity Extraction.ipynb\" target='_blank'>\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",
@@ -1428,7 +1428,7 @@
],
"metadata": {
"colab": {
"name": "sdk-automl-text-entity-extraction-batch-prediction.ipynb",
"name": "UJ7 Vertex SDK AutoML Text Entity Extraction.ipynb",
"toc_visible": true
},
"kernelspec": {
@@ -1413,7 +1413,7 @@
],
"metadata": {
"colab": {
"name": "sdk-automl-text-sentiment-analysis-batch-prediction.ipynb",
"name": "UJ8 Vertex SDK AutoML Text Sentiment Analysis.ipynb",
"toc_visible": true
},
"kernelspec": {
@@ -33,18 +33,18 @@
"\n",
"<table align=\"left\">\n",
" <td>\n",
"<a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-custom-xgboost-prebuilt-container.ipynb\" target='_blank'>\n",
"<a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ9 Vertex SDK Custom XGBoost with pre-built training container.ipynb\" target='_blank'>\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/migration/sdk-custom-xgboost-prebuilt-container.ipynb\" target='_blank'>\n",
"<a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ9 Vertex SDK Custom XGBoost with pre-built training container.ipynb\" target='_blank'>\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/migration/sdk-custom-xgboost-prebuilt-container.ipynb\" target='_blank'>\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/migration/UJ9 Vertex SDK Custom XGBoost with pre-built training container.ipynb\" target='_blank'>\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",
@@ -1459,7 +1459,7 @@
],
"metadata": {
"colab": {
"name": "sdk-custom-xgboost-prebuilt-container.ipynb",
"name": "UJ9 Vertex SDK Custom XGBoost with pre-built training container.ipynb",
"toc_visible": true
},
"kernelspec": {
@@ -33,18 +33,18 @@
"\n",
"<table align=\"left\">\n",
" <td>\n",
"<a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/sdk-custom-scikit-learn-prebuilt-container.ipynb\" target='_blank'>\n",
"<a href=\"https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ10 Vertex SDK Custom Scikit-Learn with pre-built training container.ipynb\" target='_blank'>\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/migration/sdk-custom-scikit-learn-prebuilt-container.ipynb\" target='_blank'>\n",
"<a href=\"https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/migration/UJ10 Vertex SDK Custom Scikit-Learn with pre-built training container.ipynb\" target='_blank'>\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/migration/sdk-custom-scikit-learn-prebuilt-container.ipynb\" target='_blank'>\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/migration/UJ10 Vertex SDK Custom Scikit-Learn with pre-built training container.ipynb\" target='_blank'>\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",
@@ -50,7 +50,7 @@ Learn how to use `Vertex AI Pipelines` and `Google Cloud Pipeline Components` to
The steps performed include:
- Create a Vertex AI `Dataset`.
- Train a Automl Text Classification model on the `Dataset` resource.
- Train a Automl Tabular Classification model on the `Dataset` resource.
- Import the trained `AutoML model resource` into the pipeline.
- Run a `Batch Prediction` job.
- Evaulate the AutoML model using the `Classification Evaluation Component`.
@@ -152,5 +152,3 @@ The steps performed include:
```
&nbsp;&nbsp;&nbsp;Learn more about [Model evaluation in Vertex AI](https://cloud.google.com/vertex-ai/docs/evaluation/introduction).
+2 -69
View File
@@ -23,7 +23,7 @@ The steps performed include:
&nbsp;&nbsp;&nbsp;Learn more about [Classification for tabular data](https://cloud.google.com/vertex-ai/docs/tabular-data/classification-regression/overview).
[Challenger vs Blessed methodology for model deployment into production](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/pipelines/challenger_vs_blessed_deployment_method.ipynb)
[Challenger vs Blessed methodology for model deployment into production](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/pipelines/challenger_vs_blessed_deployment_method.ipynb)
```
Learn how to construct a Vertex AI pipeline, which trains a new challenger version of a model, evaluates the model and compares the evaluation to the existing blessed model in production, to determine whether the challenger model becomes the blessed model for replacement in production.
@@ -45,10 +45,6 @@ The steps performed include:
```
&nbsp;&nbsp;&nbsp;Learn more about [Vertex AI Pipelines](https://cloud.google.com/vertex-ai/docs/pipelines/introduction).
&nbsp;&nbsp;&nbsp;Learn more about [Model evaluation in Vertex AI](https://cloud.google.com/vertex-ai/docs/evaluation/introduction).
[Pipeline control structures using the KFP SDK](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/pipelines/control_flow_kfp.ipynb)
@@ -88,7 +84,7 @@ The steps performed include:
&nbsp;&nbsp;&nbsp;Learn more about [Custom training components](https://cloud.google.com/vertex-ai/docs/training/create-training-pipeline).
[Training and batch prediction with BigQuery source and destination for a custom tabular classification model](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/pipelines/custom_tabular_train_batch_pred_bq_pipeline.ipynb)
[Training and batch prediction with BigQuery source and destinantion for a custom tabular classification model](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/pipelines/custom_tabular_train_batch_pred_bq_pipeline.ipynb)
```
In this tutorial, you train a scikit-learn tabular classification model and create batch prediction job for it through a Vertex AI pipeline using `google_cloud_pipeline_components`.
@@ -114,43 +110,6 @@ The steps performed include:
&nbsp;&nbsp;&nbsp;Learn more about [Vertex AI Batch Prediction components](https://cloud.google.com/vertex-ai/docs/pipelines/batchprediction-component).
[Get started with Vertex AI Hyperparameter Tuning pipeline components](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/pipelines/get_started_with_hpt_pipeline_components.ipynb)
```
Learn how to use prebuilt `Google Cloud Pipeline Components` for `Vertex AI Hyperparameter Tuning`.
The steps performed include:
- Construct a pipeline for:
- Hyperparameter tune/train a custom model.
- Retrieve the tuned hyperparameter values and metrics to optimize.
- If the metrics exceed a specified threshold.
- Get the location of the model artifacts for the best tuned model.
- Upload the model artifacts to a `Vertex AI Model` resource.
- Execute a Vertex AI pipeline.
```
&nbsp;&nbsp;&nbsp;Learn more about [Vertex AI Pipelines](https://cloud.google.com/vertex-ai/docs/pipelines/introduction).
&nbsp;&nbsp;&nbsp;Learn more about [Vertex AI Hyperparameter Tuning](https://cloud.google.com/vertex-ai/docs/training/hyperparameter-tuning-overview).
[Get started with machine management for Vertex AI Pipelines](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/pipelines/get_started_with_machine_management.ipynb)
```
Learn how to convert a self-contained custom training component into a `Vertex AI CustomJob`, whereby:
The steps performed in this tutorial include:
- Create a custom component with a self-contained training job.
- Execute pipeline using component-level settings for machine resources
- Convert the self-contained training component into a `Vertex AI CustomJob`.
- Execute pipeline using customjob-level settings for machine resources
```
[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)
```
@@ -339,32 +298,6 @@ The steps performed include:
&nbsp;&nbsp;&nbsp;Learn more about [Vertex AI Pipelines](https://cloud.google.com/vertex-ai/docs/pipelines/introduction).
[Multicontender vs Champion methodology for model deployment into production](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/pipelines/multicontender_vs_champion_deployment_method.ipynb)
```
Learn how to construct a Vertex AI pipeline, which evaluates new production data from a deployed model against other versions of the model, to determine if a contender model becomes the champion model for replacement in production.
The steps performed include:
- Import a pretrained (champion) model to the `Vertex AI Model Registry`.
- Import synthetic model training evaluation metrics to the corresponding (champion) model.
- Create a `Vertex AI Endpoint` resource
- Deploy the champion model to the `Endpoint` resource.
- Import additional (contender) versions of the deployed model.
- Import synthetic model training evaluation metrics to the corresponding (contender) models.
- Create a Vertex AI Pipeline
- Get the champion model.
- (Fake) Fine-tune champion model with production data
- Import synthetic train+production evaluation metrics for the champion model.
- Get the contender models.
- (Fake) Fine-tune contender model with production data
- Import synthetic train+production evaluation metrics for the contenders modesl.
- Compare the evaluations of the contenders to the champion and set the new champion as the default.
- Deploy the new champion model.
```
[Pipelines introduction for KFP](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/pipelines/pipelines_intro_kfp.ipynb)
```
@@ -265,7 +265,7 @@
"source": [
"# Install Python package dependencies.\n",
"print(\"Installing libraries\")\n",
"! pip3 install {USER_FLAG} --quiet 'google-cloud-pipeline-components==1.0.20' 'kfp<2'\n",
"! pip3 install {USER_FLAG} --quiet google-cloud-pipeline-components kfp\n",
"! pip3 install {USER_FLAG} --quiet --upgrade google-cloud-aiplatform google-cloud-bigquery"
]
},
-37
View File
@@ -14,40 +14,3 @@ The steps performed include:
&nbsp;&nbsp;&nbsp;Learn more about [Vertex AI Batch Prediction](https://cloud.google.com/vertex-ai/docs/tabular-data/classification-regression/get-batch-predictions).
[Serving PyTorch image models with prebuilt containers on Vertex AI](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/prediction/pytorch_image_classification_with_prebuilt_serving_containers.ipynb)
```
Learn how to package and deploy a PyTorch image classification model using a prebuilt Vertex AI container with TorchServe for serving online and batch predictions.
The steps performed include:
- Download a pretrained image model from PyTorch
- Create a custom model handler
- Package model artifacts in a model archive file
- Upload model for deployment
- Deploy model for prediction
- Make online predictions
- Make batch predictions
```
&nbsp;&nbsp;&nbsp;Learn more about [Pre-built containers for prediction](https://cloud.google.com/vertex-ai/docs/predictions/pre-built-containers).
[Train and deploy PyTorch models with prebuilt containers on Vertex AI](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/prediction/pytorch_train_deploy_models_with_prebuilt_containers.ipynb)
```
Learn how to build, train and deploy a PyTorch image classification model using prebuilt containers for custom training and prediction.
The steps performed include:
- Package training application into a Python source distribution
- Configure and run training job in a prebuilt container
- Package model artifacts in a model archive file
- Upload model for deployment
- Deploy model using a prebuilt container for prediction
- Make online predictions
```
-20
View File
@@ -15,26 +15,6 @@ The steps performed are:
&nbsp;&nbsp;&nbsp;Learn more about [Tabular Workflow for TabNet](https://cloud.google.com/vertex-ai/docs/tabular-data/tabular-workflows/tabnet).
[Get started with TabNet builtin algorithm for training tabular models](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/tabnet/get_started_with_tabnet.ipynb)
```
Learn how to run `Vertex AI TabNet` built algorithm for training custom tabular models.
The steps performed include:
- Get the training data.
- Configure training parameters for the `Vertex AI TabNet` container.
- Train the model using `Vertex AI Training` using CSV data.
- Upload the model as a `Vertex AI Model` resource.
- Deploy the `Vertex AI Model` resource to a `Vertex AI Endpoint` resource.
- Make a prediction with the deployed model.
- Hyperparameter tuning the `Vertex AI TabNet` model.
```
&nbsp;&nbsp;&nbsp;Learn more about [Tabular Workflow for TabNet](https://cloud.google.com/vertex-ai/docs/tabular-data/tabular-workflows/tabnet).
[Vertex AI TabNet](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/tabnet/tabnet_vertex_tutorial.ipynb)
```
@@ -14,8 +14,6 @@ The steps performed are:
&nbsp;&nbsp;&nbsp;Learn more about [Google Cloud Pipeline Components](https://cloud.google.com/vertex-ai/docs/pipelines/components-introduction).
&nbsp;&nbsp;&nbsp;Learn more about [Prophet for tabular data](https://cloud.google.com/vertex-ai/docs/tabular-data/forecasting-prophet).
[TabNet Pipeline](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/tabular_workflows/tabnet_on_vertex_pipelines.ipynb)
-16
View File
@@ -10,19 +10,3 @@ Learn how to use `Vertex AI Vizier` to optimize a multi-objective study.
&nbsp;&nbsp;&nbsp;Learn more about [Vertex AI Vizier](https://cloud.google.com/vertex-ai/docs/vizier/overview).
[Get started with Vertex AI Vizier](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/vizier/get_started_vertex_vizier.ipynb)
```
Learn how to use `Vertex AI Vizier` for when training with `Vertex AI`.
The steps performed include:
- Hyperparameter tuning with Random algorithm.
- Hyperparameter tuning with Vizier (Bayesian) algorithm.
- Suggesting trials and updating results for Vizier study
```
&nbsp;&nbsp;&nbsp;Learn more about [Vertex AI Vizier](https://cloud.google.com/vertex-ai/docs/vizier/overview).
+4 -4
View File
@@ -81,7 +81,7 @@ The steps performed include:
```
&nbsp;&nbsp;&nbsp;Learn more about [Vertex AI Workbench](https://cloud.google.com/vertex-ai/docs/workbench/managed/visualize-data-bigquery).
&nbsp;&nbsp;&nbsp;Learn more about [Vertex AI Workbench](https://cloud.google.com/vertex-ai/docs/workbench/introduction).
&nbsp;&nbsp;&nbsp;Learn more about [BigQuery ML](https://cloud.google.com/vertex-ai/docs/beginner/bqml).
@@ -127,7 +127,7 @@ The steps performed include:
&nbsp;&nbsp;&nbsp;Learn more about [Vertex AI Workbench](https://cloud.google.com/vertex-ai/docs/workbench/introduction).
&nbsp;&nbsp;&nbsp;Learn more about [BigQuery ML](https://cloud.google.com/vertex-ai/docs/beginner/bqml#machine_learning_directly_in).
&nbsp;&nbsp;&nbsp;Learn more about [BigQuery ML](https://cloud.google.com/vertex-ai/docs/beginner/bqml).
[Inventory prediction on ecommerce data using Vertex AI](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/workbench/inventory-prediction/inventory_prediction.ipynb)
@@ -194,7 +194,7 @@ The steps performed include:
&nbsp;&nbsp;&nbsp;Learn more about [Vertex AI Workbench](https://cloud.google.com/vertex-ai/docs/workbench/introduction).
&nbsp;&nbsp;&nbsp;Learn more about [BigQuery ML](https://cloud.google.com/vertex-ai/docs/beginner/bqml#machine_learning_directly_in).
&nbsp;&nbsp;&nbsp;Learn more about [BigQuery ML](https://cloud.google.com/vertex-ai/docs/beginner/bqml).
[Sentiment Analysis using AutoML Natural Language and Vertex AI](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/workbench/sentiment_analysis/Sentiment_Analysis.ipynb)
@@ -218,7 +218,7 @@ The steps performed are:
&nbsp;&nbsp;&nbsp;Learn more about [Vertex AI Workbench](https://cloud.google.com/vertex-ai/docs/workbench/introduction).
&nbsp;&nbsp;&nbsp;Learn more about [Sentiment analysis for text data](https://cloud.google.com/vertex-ai/docs/training-overview#sentiment_analysis_for_text).
&nbsp;&nbsp;&nbsp;Learn more about [AutoML Text](https://cloud.google.com/vertex-ai/docs/tutorials/text-classification-automl/training).
[Digest and analyze data from BigQuery with Dataproc](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/workbench/spark/spark_bigquery.ipynb)