NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
PyPI · #1039 most downloaded on PyPI
Open source library for training and deploying models on Amazon SageMaker.
Last release today
24 Sep 2026
Ships on a steady schedule
a new release about every 2 weeks
Nearly every release is documented
notes for 59 of the last 60 stable releases
3 versions withdrawn
withdrawn after publishing
9 years old
692 releases · first in 2017
repack_model support dependencies and code location
add region check for Neo service
One column per quarter.
update using_sklearn.rst parameter name
add encryption option to "record_set"
set _current_job_name in attach()
repack model function works without source directory
### Features * Support for TFS preprocessing
run tests if buildspec.yml has been modified
pin pytest version to 4.4.1 to avoid pluggy version conflict
update TrainingInputMode with s3_input InputMode
prevent false positive PR test results
add py2 deprecation message for the deep learning framework images
add document embedding support to Object2Vec algorithm
skip p2/p3 tests in eu-central-1
add automatic model tuning integ test for TF script mode
use unique names for test training jobs
add KMS key option for Endpoint Configs
fix propagation of tags to SageMaker endpoint
remove duplicate content from Chainer readme
remove duplicate content from PyTorch readme and fix internal links
make Local Mode export artifacts even after failure
skip horovod p3 test in region with no p3
use unique names for test training jobs
local data source relative path includes the first directory
make start time, end time and period configurable in sagemaker.analytics.TrainingJobAnalytics
### Documentation changes * spelling error correction
move RL readme content into sphinx project
hyperparameter query failure on script mode estimator attached to complete job
remove unrestrictive principal * from KMS policy tests.
add pytest marks for integ tests using local mode
add mandatory sagemaker_role argument to Local mode example.
bug-fix: pass kms id as parameter for uploading code with Server side encryption
PipelineModel: Create a Transformer from a PipelineModelAlgorithmEstimator: Make SupportedHyperParameters optionalHyperparameter: Support scaling hyperparametersdoc-fix: Remove incorrect parameter for EI TFS Python README
Predictor: delete SageMaker modelPipelineModel: delete SageMaker modeldeploy for REST API TFS Model* doc-fix: fix README for PyPI
doc-fix: update information about saving models in the MXNet README
transform_fn information and fix input_fn signature in the MXNet READMEPredictor to delete endpoint configuration by default when calling delete_endpoint()Model to delete SageMaker modelTransformer to delete SageMaker modelenhancement: Include SageMaker Notebook Instance version number in boto3 user agent, if available.
enhancement: Add tuner to imports in sagemaker/__init__.py
tuner to imports in sagemaker/__init__.pybug-fix: Handle StopIteration in CloudWatch Logs retrieval
feature: HyperparameterTuner: support VPC config
enhancement: Workflow: Specify tasks from which training/tuning operator to transform/deploy in related operators
bug-fix: Workflow: Revert appending Airflow retry id to default job name
__all__ from __init__.py filesSM_HPS environment variable in MXNet READMEinclude_cls_metadata default to False for everything except Frameworksbug-fix: Local Mode: Allow support for SSH in local mode
enhancement: Check for S3 paths being passed as entry point
distribution to distributionsDocumentation: add documentation for Reinforcement Learning Estimator.
feature: update boto3 to version 1.9.55
bug-fix: Fix FileNotFoundError for entry_point without source_dir
enhancement: Local Mode: add explicit pull for serving
feature: Estimator: add script mode and Python 3 support for TensorFlow
enhancement: Frameworks: update warning for not setting framework_version as we aren't planning a breaking change anymore
CustomAttributes argument in local mode invoke_endpoint requestscontent_type parameter to sagemaker.tensorflow.serving.Predictorfeature: Estimators: add support for Amazon Object2Vec algorithm
feature: add support for sagemaker-tensorflow-serving container
feature: Estimator: add input mode to training channels
distributions for customizing distributed training with the new training script formatfeature: add support for TensorFlow 1.11.0
warning: Frameworks: add warning for upcoming breaking change that makes framework_version required
enhancement: Enable setting VPC config when creating/deploying models
enhancement: Local Mode: add training environment variables for AWS region and job name
bug-fix: setting health check timeout limit on local mode to 30s
doc-fix: add deprecation warning for current MXNet training script format
feature: add support for TensorFlow 1.10.0
doc-fix: fix rst warnings in README.rst
Your coding agent can read these notes before it upgrades. Set up the MCP server →