2 Pull requests merged by 2 people

• fix(webdataset): don't .lower() field_name

  #7726 merged Aug 20, 2025

• HDF5 support

  #7690 merged Aug 19, 2025

4 Pull requests opened by 2 people

• Document HDF5 support

  #7740 opened Aug 19, 2025

• Refactor HDF5 and preserve tree structure

  #7743 opened Aug 21, 2025

• Add wikipedia-2023-redirects dataset

  #7747 opened Aug 22, 2025

• docs: Streaming best practices

  #7748 opened Aug 23, 2025

1 Issue closed by 1 person

• Loading Data from HDF files

  #3113 closed Aug 19, 2025

7 Issues opened by 7 people

• Fix: Canonical 'multi_news' dataset is broken and should be updated to a Parquet version

  #7746 opened Aug 22, 2025

• Audio mono argument no longer supported, despite class documentation

  #7745 opened Aug 22, 2025

• dtype: ClassLabel is not parsed correctly in `features.py`

  #7744 opened Aug 21, 2025

• module 'pyarrow' has no attribute 'PyExtensionType'

  #7742 opened Aug 20, 2025

• Preserve tree structure when loading HDF5

  #7741 opened Aug 19, 2025

• Replacement of "Sequence" feature with "List" breaks backward compatibility

  #7739 opened Aug 18, 2025

• Allow saving multi-dimensional ndarray with dynamic shapes

  #7738 opened Aug 18, 2025

3 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.

• Add columns support to JSON loader for selective key filtering

  #7652 commented on Aug 18, 2025 • 1 new comment

• Dataset scripts are no longer supported, but found superb.py

  #7693 commented on Aug 20, 2025 • 0 new comments

• Add the possibility of a backend for audio decoding

  #7731 commented on Aug 20, 2025 • 0 new comments