Releases: aws/sagemaker-python-sdk
Releases · aws/sagemaker-python-sdk
v2.183.0
Deprecations and Removals
- remove support for py37
Features
- Neo service GA in TLV
Bug Fixes and Other Changes
- Update pytorch.json with 2.0.1 for inference and training
- get python version dynamically for remote function tests
- HuggingFaceProcessor parameterized instance_type when image_uri is absent
v2.182.0
Features
- image url for modelmonitor in TLV region
- Enable spot training on remote decorator and executor
v2.181.0
Features
- StabilityAI DLC Image URIs
Bug Fixes and Other Changes
- temporarily skip kmeans notebook
v2.180.0
Features
- Add presigned URLs for interactive apps
- Add detail profiler V2 options and tests
v2.179.0
Features
- attach method for jumpstart estimator
Bug Fixes and Other Changes
- pipeline upsert failed to pass parallelism_config to update
v2.178.0
Features
- Support to get latest monitoring execution processing logs
Bug Fixes and Other Changes
- Add context to predict_fn example
- gated models unsupported region
- jumpstart cache using sagemaker session s3 client
- add TFS 2.13 Graviton SM images
- pipeline variable kms key
- integration test for gated jumpstart training model
- tags for jumpstart model package models
v2.177.1
Bug Fixes and Other Changes
- chore: excessive jumpstart bucket logging
v2.177.0
Features
- Add TLV accounts for 1P Algorithms
v2.176.0
Features
- Add TF 2.13 Training and Inference SM images
Bug Fixes and Other Changes
- revert-PR_3903
- skip tensorflow local mode notebook test
- change instance type for huggingface test to ml.g5.8xlarge
v2.175.0
Features
- Add huggingface-llm 0.9.3 dlc images
Bug Fixes and Other Changes
- Upgrade default version for djl to v0.23.0
- Pass kms_key to _upload_analysis_config when provided