Releases: aws/sagemaker-python-sdk
Releases · aws/sagemaker-python-sdk
SageMaker Python SDK 1.4.2
- bug-fix: Unit Tests: Improve unit test runtime
- bug-fix: Estimators: Fix attach for LDA
- bug-fix: Estimators: allow code_location to have no key prefix
- bug-fix: Local Mode: Fix s3 training data download when there is a trailing slash
SageMaker Python SDK 1.4.1
- bug-fix: Local Mode: Fix for non Framework containers
SageMaker Python SDK 1.4.0
- bug-fix: Remove all and add noqa in init
- bug-fix: Estimators: Change max_iterations hyperparameter key for KMeans
- bug-fix: Estimators: Remove unused argument job_details for
EstimatorBase.attach()
- bug-fix: Local Mode: Show logs in Jupyter notebooks
- feature: HyperparameterTuner: Add support for hyperparameter tuning jobs
- feature: Analytics: Add functions for metrics in Training and Hyperparameter Tuning jobs
- feature: Estimators: add support for tagging training jobs
SageMaker Python SDK 1.3.0
- feature: add chainer
SageMaker Python SDK 1.2.5
- bug-fix: Change module names to string type in all
- feature: Save training output files in local mode
- bug-fix: tensorflow-serving-api: SageMaker does not conflict with tensorflow-serving-api module version
- feature: Local Mode: add support for local training data using file://
- feature: Updated TensorFlow Serving api protobuf files
- bug-fix: No longer poll for logs from stopped training jobs
SageMaker Python SDK 1.2.4
- feature: Estimators: add support for Amazon Random Cut Forest algorithm
SageMaker Python SDK 1.2.3
1.2.3
- bug-fix: Fix local mode not using the right s3 bucket
SageMaker Python SDK v1.2.2
1.2.2
- bug-fix: Estimators: fix valid range of hyper-parameter 'loss' in linear learner
SageMaker Python SDK 1.2.1
1.2.1
- bug-fix: Change Local Mode to use a sagemaker-local docker network
v1.2.0: Update default versions to TensorFlow 1.6 and MXNet 1.1 (#118)
1.2.0
- feature: Add Support for Local Mode
- feature: Estimators: add support for TensorFlow 1.6.0
- feature: Estimators: add support for MXNet 1.1.0
- feature: Frameworks: Use more idiomatic ECR repository naming scheme