v4.1.0

Dataset Features

• feat: use content defined chunking by @kszucs in https://github.com/huggingface/datasets/pull/7589

  • Parquet datasets are now Optimized Parquet !

    <img width="462" height="103" alt="image" src="https://github.com/user-attachments/assets/43703a47-0964-421b-8f01-1a790305de79" />

  • internally uses use_content_defined_chunking=True when writing Parquet files

  • this enables fast deduped uploads to Hugging Face !

  # Now faster thanks to content defined chunking
  ds.push_to_hub("username/dataset_name")

  • this optimizes Parquet for Xet, the dedupe-based storage backend of Hugging Face. It allows to not have to upload data that already exist somewhere on HF (on an other file / version for example). Parquet content defined chunking defines Parquet pages boundaries based on the content of the data, in order to detect duplicate data easily.
  • with this change, the new default row group size for Parquet is set to 100MB
  • write_page_index=True is also used to enable fast random access for the Dataset Viewer and tools that need it

• Concurrent push_to_hub by @lhoestq in https://github.com/huggingface/datasets/pull/7708

• Concurrent IterableDataset push_to_hub by @lhoestq in https://github.com/huggingface/datasets/pull/7710

• HDF5 support by @klamike in https://github.com/huggingface/datasets/pull/7690

  • load HDF5 datasets in one line of code

  ds = load_dataset("username/dataset-with-hdf5-files")

  • each (possibly nested) field in the HDF5 file is parsed a a column, with the first dimension used for rows

Other improvements and bug fixes

• Convert to string when needed + faster .zstd by @lhoestq in https://github.com/huggingface/datasets/pull/7683
• fix audio cast storage from array + sampling_rate by @lhoestq in https://github.com/huggingface/datasets/pull/7684
• Fix misleading add_column() usage example in docstring by @ArjunJagdale in https://github.com/huggingface/datasets/pull/7648
• Allow dataset row indexing with np.int types (#7423) by @DavidRConnell in https://github.com/huggingface/datasets/pull/7438
• Update fsspec max version to current release 2025.7.0 by @rootAvish in https://github.com/huggingface/datasets/pull/7701
• Update dataset_dict push_to_hub by @lhoestq in https://github.com/huggingface/datasets/pull/7711
• Retry intermediate commits too by @lhoestq in https://github.com/huggingface/datasets/pull/7712
• num_proc=0 behave like None, num_proc=1 uses one worker (not main process) and clarify num_proc documentation by @tanuj-rai in https://github.com/huggingface/datasets/pull/7702
• Update cli.mdx to refer to the new "hf" CLI by @evalstate in https://github.com/huggingface/datasets/pull/7713
• fix num_proc=1 ci test by @lhoestq in https://github.com/huggingface/datasets/pull/7714
• Docs: Use Image(mode="F") for PNG/JPEG depth maps by @lhoestq in https://github.com/huggingface/datasets/pull/7715
• typo by @lhoestq in https://github.com/huggingface/datasets/pull/7716
• fix largelist repr by @lhoestq in https://github.com/huggingface/datasets/pull/7735
• Grammar fix: correct "showed" to "shown" in fingerprint.py by @brchristian in https://github.com/huggingface/datasets/pull/7730
• Fix type hint train_test_split by @qgallouedec in https://github.com/huggingface/datasets/pull/7736
• fix(webdataset): don't .lower() field_name by @YassineYousfi in https://github.com/huggingface/datasets/pull/7726
• Refactor HDF5 and preserve tree structure by @klamike in https://github.com/huggingface/datasets/pull/7743
• docs: Add column overwrite example to batch mapping guide by @Sanjaykumar030 in https://github.com/huggingface/datasets/pull/7737
• Audio: use TorchCodec instead of Soundfile for encoding by @lhoestq in https://github.com/huggingface/datasets/pull/7761
• Support pathlib.Path for feature input by @Joshua-Chin in https://github.com/huggingface/datasets/pull/7755
• add support for pyarrow string view in features by @onursatici in https://github.com/huggingface/datasets/pull/7718
• Fix typo in error message for cache directory deletion by @brchristian in https://github.com/huggingface/datasets/pull/7749
• update torchcodec in ci by @lhoestq in https://github.com/huggingface/datasets/pull/7764
• Bump dill to 0.4.0 by @Bomme in https://github.com/huggingface/datasets/pull/7763

New Contributors

• @DavidRConnell made their first contribution in https://github.com/huggingface/datasets/pull/7438
• @rootAvish made their first contribution in https://github.com/huggingface/datasets/pull/7701
• @tanuj-rai made their first contribution in https://github.com/huggingface/datasets/pull/7702
• @evalstate made their first contribution in https://github.com/huggingface/datasets/pull/7713
• @brchristian made their first contribution in https://github.com/huggingface/datasets/pull/7730
• @klamike made their first contribution in https://github.com/huggingface/datasets/pull/7690
• @YassineYousfi made their first contribution in https://github.com/huggingface/datasets/pull/7726
• @Sanjaykumar030 made their first contribution in https://github.com/huggingface/datasets/pull/7737
• @kszucs made their first contribution in https://github.com/huggingface/datasets/pull/7589
• @Joshua-Chin made their first contribution in https://github.com/huggingface/datasets/pull/7755
• @onursatici made their first contribution in https://github.com/huggingface/datasets/pull/7718
• @Bomme made their first contribution in https://github.com/huggingface/datasets/pull/7763

Full Changelog: https://github.com/huggingface/datasets/compare/4.0.0...4.1.0