All pull requests

Pull requests

• Fix DatasetDict examples in the NumPy and JAX guides

  #8783·shaikn6 opened Oct 11, 2026·

  ·

• perf: speed up dense indexed batch reads

  #8782·n33levo opened Oct 11, 2026·

  ·

• Fix serialization of mesh objects returned by Dataset.map

  #8781·antonsoo opened Oct 10, 2026·

  ·

• Accept LargeList embedding columns in FAISS indexes

  #8780·ktz03 opened Oct 10, 2026·

  ·

• Close CSV readers when table generation exits

  #8779·Nikillanr opened Oct 10, 2026·

  ·

• Keep CSV row indices continuous across batches

  #8778·j45856021-dev opened Oct 9, 2026·

  ·

• Make Features.__setitem__/__delitem__ update only the touched column

  #8777·novain-shibu opened Oct 9, 2026·

  ·
  1

• fix: close parquet writer even when export fails (#8743)

  #8775·amengnew opened Oct 9, 2026·

  ·
  1

• Fix row-wise equality for pandas ArrayXD columns

  #8774·ktz03 opened Oct 9, 2026·

  ·

• Add return_file_name parameter to load_dataset for JSON (issue #5806)

  #8773·shivanshutiwari35 opened Oct 8, 2026·

  ·
  1

• Use Sequence instead of List in type hints (issue #5354)

  #8772·shivanshutiwari35 opened Oct 8, 2026·

  ·
  1

• Honor NumPy dtype and copy in ArrayXD adapters

  #8771·Afloat16 opened Oct 8, 2026·

  ·

• Fix load_from_disk failing for datasets saved with zero rows

  #8770·yupengtang opened Oct 8, 2026Contributor·

  ·

• Fix Video feature serialization with torch.device inputs

  #8769·aalquwayfili opened Oct 8, 2026·

  ·

• Fix Arrow iterable map preserving returned columns

  #8768·97Three opened Oct 8, 2026·

  ·
  1

• Fix struct-backed feature alignment in Dataset.add_item

  #8767·LiamLin1130 opened Oct 7, 2026·

  ·
  1

• Infer WebDataset media features through compression suffixes

  #8765·parlorsky opened Oct 7, 2026·

  ·

• Fix IndexError for an empty contiguous selection in Dataset.select and shard

  #8764·Mudassiruddin7 opened Oct 7, 2026·

  ·
  1

  2

• Decode the uint32 TenBin type code in WebDataset

  #8762·LeonSGP43 opened Oct 7, 2026·

  ·
  1

• Reject unequal row counts in selected nested HDF5 datasets

  #8761·LeonSGP43 opened Oct 7, 2026·

  ·

• Distinguish RIFF WAVE audio from AVI video when loading Lance

  #8760·LeonSGP43 opened Oct 7, 2026·

  ·

• Include Iceberg catalog properties in the dataset cache identity

  #8759·LeonSGP43 opened Oct 7, 2026·

  ·

• Fix serialization of in-memory PDF objects without moving their stream

  #8758·LeonSGP43 opened Oct 7, 2026·

  ·

• Fix audio examples in the format docs

  #8757·jayzuccarelli opened Oct 6, 2026Contributor·

  ·
  2

• Support TensorFlow conversion of Dataset columns

  #8756·gauravch-code opened Oct 6, 2026·

  ·
  1