* Update FeatureStore embedding notebook to use vertex SDK preview.
* Fix format issue.
* Comment deleting transfer config code
* Add Credentialed Accounts
ACTIVE ACCOUNT
* 141951627079-compute@developer.gserviceaccount.com to debug permission issue
* Fix bigquery transfer run issue.
* Initial commit of a Llama2 TPUv5e LoRA example
* Addressing issue of job failing at the end by adding sys.exit(0) after all steps complete
* Adding Llama2 LoRA tuning on TPUv5e notebook
* Create /tmp/modelfiles folder for downloading model files during training
* Updated to use gcloud storage instead of gsutil
* Removed step to clear folder in bucket, since that causes failure
* Added serving/deployment section.
* Fixed linting errors
* Default to 8 chip tuning for quota issue on automated tests
* Defaulting back to 16 chip fine-tuning and epochs to 200
* Set Accelerate library to 0.28.0, as newer version break TPU support
* Updated upload of tuned model files to happen for all workers - to avoid timeout errors
* Temp remove active endpoint, then retest
* Updating docker container to python 3.10
* Added wait to prevent job delete step happening too early
* Lint issue and remove temporary endpoint delete
* Lint issue
* Add 15 minute wait while model is setup on the endpoint, and readd temp delete of endpoint
* Make 15 minute wait optional and remove temporary endpoint delete
* Remove GCS path to model garden model for final commit
---------
Co-authored-by: Rob Vogelbacher <robv@google.com>
* Update FS optimized serving GA colab
* Fix import order
* Fix import format
* Fix format
* Fix test issue
* Fix format
* Add way to getFOS&FV for FR/FG
* Fix minor typo in accelerator selection.
* Fix minor typo in accelerator selection.
* Update eval dataset to hellaswag.
* Update eval dataset to hellaswag.
* feat: Update supported models table in RLHF notebooks.
* feat: Update image that shows how to locate output_model_path.
* feat: Update name of deploy model component.
---------
Co-authored-by: Ryan Latture <latture@google.com>
* Add Colab on how to use Claude 3 models on Vertex AI
* Add Colab on how to use Claude 3 models on Vertex AI
* add codeowner
* lint
* update model version
* update stream sdk to have more readable response
* add colab for claude 3
* lint
* update CODEOWNERS and file name
* remove claude 3 colab from model_garden folder
* update CODEOWNERS
* format
* fix intergration test failed
* fix markdown not showing up
* fix markdown not showing up
* update pip command
* update install command
* update restart kernel command
* fix restart kernel command
* clear cell output
* update select region command
* raise error if user doesn't update project_id
* update ordering and add preview image section
* update ordering and add preview image section
* fix httpx package not install
* fix httpx package not install
* fix httpx package not install
* fix lint
* add claude 3 opus model
* fix lint
---------
Co-authored-by: Huy Ngo <huyngo@google.com>
* Add Colab on how to use Claude 3 models on Vertex AI
* Add Colab on how to use Claude 3 models on Vertex AI
* add codeowner
* lint
* update model version
* update stream sdk to have more readable response
* add colab for claude 3
* lint
* update CODEOWNERS and file name
* remove claude 3 colab from model_garden folder
* update CODEOWNERS
* format
* fix intergration test failed
* fix markdown not showing up
* fix markdown not showing up
* update pip command
* update install command
* update restart kernel command
* fix restart kernel command
* clear cell output
* update select region command
* raise error if user doesn't update project_id
* update ordering and add preview image section
* update ordering and add preview image section
* fix httpx package not install
* fix httpx package not install
* fix httpx package not install
* fix lint
---------
Co-authored-by: Huy Ngo <huyngo@google.com>
* Some minor updates to the SD-XL model deployment notebook, based on the QA feedback.
* Updates to the `sd-xl-dreambooth-lora-finetune` notebook based on feedback.
* [Stable diffusion gradio] Add a few pre defined styles to the workshop, also some minor UX improvement.
* Update all the diffusion-serve containder URI to 20240403_0836_RC00 which includes the latest optimizations to the diffusion models.
* Some minor updates to the SD-XL model deployment notebook, based on the QA feedback.
* Updates to the `sd-xl-dreambooth-lora-finetune` notebook based on feedback.
* [Stable diffusion gradio] Add a few pre defined styles to the workshop, also some minor UX improvement.
* Some minor updates to the SD-XL model deployment notebook, based on the QA feedback.
* Updates to the `sd-xl-dreambooth-lora-finetune` notebook based on feedback.
