aws / aws/sagemaker-python-sdk
Allow sagemaker.clarify.ModelConfig to use application/x-parquet for content_type & accept_type
- Dominant language
- Python
- Stars
- 2.3k
- Forks
- 1.3k
- Avg merge
- 2d 2h
- Merged PRs (30d)
- 32
Description
**Describe the feature you'd like**
Allow `sagemaker.clarify.ModelConfig` object to accept `application/x-parquet` as a valid parameter for `content_type` & `accept_type`.
**How would this feature be used? Please describe.**
This would align with the `sagemaker.clarify.DataConfig` object which currently allows parquet for the `dataset_type`. By implementing this feature, Bring-your-own-container inference models that only accept parquet data types will be able to more easily utilize clarify's bias and data quality features.
**Describe alternatives you've considered**
Currently, majority of our data is in parquet format. Conversion to csv is trivial, though for larger datasets becomes cumbersome.
Contributor guide
Assessment
This issue has not been assessed yet.