aws / aws/amazon-sagemaker-feedback
Support DatasetDefinition for processing job in SageMaker Pipeline's Local Mode
- Dominant language
- No language data
- Stars
- 10
- Forks
- 3
- PR merge metrics
- No merged PRs in 30d
Description
### Product Version
- [ ] Amazon SageMaker Studio Classic
- [ ] Amazon SageMaker Studio
- [X] It is not related to SageMaker Studio
### Product Category
Jobs
### Description
Local Mode is essential for quick iteration in debugging and development. The lacking support for `DatasetDefinition` basically means I can't use it, since it would break my development cycle. For example, converting the official example from [amazon-sagemaker-feature-store-end-to-end-workshop](https://github.com/aws-samples/amazon-sagemaker-feature-store-end-to-end-workshop/blob/cb083c60e7394023487f314f5392ab2c64331775/03-module-training-and-batch-scoring/m3_nb2_batch_transform.ipynb#L56) to local mode, I get the following error:
> "Step 'PrepareTransform' failed with message: RuntimeError: DatasetDefinition is not currently supported in Local Mode"
### Other Details
_No response_
Contributor guide
Research direction
Start with the linked Amazon SageMaker Feature Store workshop example and reproduce the reported Local Mode processing-job failure. Investigate the DatasetDefinition path in SageMaker Pipelines; done means a processing job using DatasetDefinition runs successfully in Local Mode without the reported unsupported error.
Written by the indexing model from the issue text.
Assessment
- Domain
- cloud, machine-learning
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100