Brings data mutation, version management, query, materialized view to Tensorflow Datasets

Hi all,

I’d like to share a new ML storage framework (GitHub - google/space: Unified storage framework for the entire machine learning lifecycle). It brings many database/lakehouse features, like data mutations, version management, SQL, materialized views of processed results, into ML datasets (currently TFDS, Ray, HuggingFace datasets are in scope).

We have a TFDS example: Feedbacks and contributions are very welcome.

Thank you!