1 Release published by 1 person

• v2.249.0

  published Jul 31, 2025

2 Pull requests merged by 2 people

• feature: AWS Batch for SageMaker Training jobs

  #5249 merged Jul 30, 2025

• Directly use customer-provided endpoint name for ModelBuilder deployment.

  #5246 merged Jul 28, 2025

1 Pull request opened by 1 person

• feat: support pipeline versioning

  #5248 opened Jul 25, 2025

9 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• Importing sagemaker seems to enable rich tracebacks

  #5179 commented on Jul 28, 2025 • 0 new comments

• Structure for custom entrypoint inference script and its overridden methods for DJLModel

  #4182 commented on Jul 28, 2025 • 0 new comments

• Something like a .sagemakerignore file option to allow for skipping of certain directories when creating source tar

  #4187 commented on Jul 28, 2025 • 0 new comments

• Document the support for installing requirements.txt dependencies from CodeArtifact in PyTorch 2.0.1+ SageMaker Containers

  #4189 commented on Jul 29, 2025 • 0 new comments

• Use of boolean hyperparameters: recommended use of `type=bool` is not okay

  #4247 commented on Jul 29, 2025 • 0 new comments

• ModelTrainer doesn't support heterogeneous clusters

  #5225 commented on Jul 30, 2025 • 0 new comments

• ModelTrainer doesn't propagate hyperparameters if SourceCode-command is used

  #5226 commented on Jul 30, 2025 • 0 new comments

• SageMaker Estimator doesn't support checkpoint_s3_uri with Heterogeneous Clusters

  #5227 commented on Jul 31, 2025 • 0 new comments

• uv add sagemaker fails on 3.13

  #5233 commented on Aug 1, 2025 • 0 new comments