sagemaker-python-sdk

Features

• support tuning step parameter range parameterization + support retry strategy in tuner

Documentation Changes

• clarify that default_bucket creates a bucket
• Minor updates to Clarify API documentation

Bug Fixes and Other Changes

• sklearn integ tests, remove swallowing exception on feature group delete attempt
• sklearn integ test for custom bucket

Documentation Changes

• Fix dataset_definition links
• Document LambdaModel and LambdaPredictor classes

Bug Fixes and Other Changes

• revert #2251 changes for sklearn processor

Features

• processors that support multiple Python files, requirements.txt, and dependencies.
• support step object in step depends on list

Bug Fixes and Other Changes

• enable isolation while creating model from job
• update sagemaker.serverless integration test
• Use correct boto model name for RegisterModelStep properties

Features

• add LambdaStep support for SageMaker Pipelines
• support JsonGet for all step types

Bug Fixes and Other Changes

• null checks for uploaded_code and entry_point

Documentation Changes

• update sagemaker.estimator.EstimatorBase
• Mark baseline as optional in KernelSHAP.

Features

• add KIX region to image_uris

Bug Fixes and Other Changes

• Rename PredictorBase.delete_endpoint as PredictorBase.delete_predictor
• incorrect default argument for callback output parameter

Documentation Changes

• Remove years from copyright boilerplate
• Fix documentation formatting for PySpark and SparkJar processors

Testing and Release Infrastructure

• enable py38 tox env

Bug Fixes and Other Changes

• order of populating container list
• upgrade Adobe Analytics cookie to 3.0

Bug Fixes and Other Changes

• Set flag when debugger is disabled
• KMS Key fix for kwargs
• Update BiasConfig to accept multiple facet params

Documentation Changes

• Update huggingface estimator documentation

Features

• Adding serial inference pipeline support to RegisterModel Step

Documentation Changes

• add tuning step get_top_model_s3_uri and callback step to doc
• links for HF in sdk
• Add Clarify module to Model Monitoring API docs

Bug Fixes and Other Changes

• default time for compilation jobs
• skip hf inference test

Bug Fixes and Other Changes

• skip HF inference test
• remove upsert from test_workflow

Documentation Changes

• Add Hugging Face docs
• add tuning step to doc

Features

• HuggingFace Inference

Bug Fixes and Other Changes

• add support for SageMaker workflow tuning step

Documentation Changes

• smddp 1.2.1 release note / convert md to rst
• add smd model parallel 1.4.0 release note / restructure doc files

Bug Fixes and Other Changes

• handle tags when upsert pipeine

Bug Fixes and Other Changes

• revert "fix: jsonGet interpolation issue 2426 + allow step depends on pass in step instance (#2477)"

Features

• support job_name_prefix for Clarify

Bug Fixes and Other Changes

• Add configuration option with headers for Clarify Explainability
• jsonGet interpolation issue 2426 + allow step depends on pass in step instance
• add default retries to feature group ingestion.
• Update using_pytorch.rst
• kms key does not propapate in register model step
• Correctly interpolate Callback output parameters

Bug Fixes and Other Changes

• Register model step tags

Documentation Changes

• update to include new batch_get_record api call
• Correct type annotation for TrainingStep inputs
• introduce input mode FastFile
• update hf transformer version

Features

• Add HF transformer version 4.6.1

Bug Fixes and Other Changes

• encode localmode payload to UTF-8
• call DescribeDomain as fallback in get_execution_role
• parameterize PT and TF version for HuggingFace tests

Documentation Changes

• Add import statement in Batch Transform Overview doc