Sagemaker

Features

* pluggable instance fallback mechanism, add CapacityError
* support passing Env Vars to local mode training

Bug Fixes and Other Changes

* Update black-check version, add support for Spark 3.1 Processing

Features

* Retrieve data configuration
* enable EnableInterContainerTrafficEncryption for model monitoring
* Hugging Face Transformers 4.17 for PT 1.10

Bug Fixes and Other Changes

* remove `new` from serverless
* temporarily skip tests impacted by data inconsistency
* Implement override solution for pipeline variables

Documentation Changes

* add documentation for image_uri serverless use case
* minor fixes for smddp 1.4.0 doc

Features

* Add support for TF2.7
* Add support for TF 2.8
* TF242 ioc support
* Add support for TF 2.6.3
* Support for remote docker host
* AutoGluon 0.3.2 and 0.4.0 image_uris

Bug Fixes and Other Changes

* Align max_wait definitions in EstimaorBase and Estimator
* Add JumpStart model table build notification
* gpu integs CapacityError - fallback to available compute
* gpu integs CapacityError - fallback to available compute
* jumpstart docs network isolation

Documentation Changes

* sagemaker distributed model parallel 1.7.0 doc

Features

* Inferentia Neuron support for HuggingFace
* custom base job name for jumpstart models/estimators
* Python 3.9 for readthedocs

Bug Fixes and Other Changes

* container env generation for S3 URI and add test for the same

Documentation Changes

* the SageMaker distributed data parallel v1.4.0 release
* update sagemaker training compiler docstring
* smddp doc update

Features

* TensorFlow 2.4 for Neo
* Data Serializer

Bug Fixes and Other Changes

* Style update in DataSerializer
* Remove sagemaker_job_name from hyperparameters in TrainingStep
* reorganize test files for workflow
* update code to get commit_id in codepipeline