ReleaseLightning AILightning AIpublished Sep 30, 2026seen 1w

Lightning-AI/litData v0.2.76

Lightning-AI/litData

Open original ↗

Captured source

source ↗
published Sep 30, 2026seen 1wcaptured 1whttp 200method plain

Release v0.2.76

Repository: Lightning-AI/litData

Tag: v0.2.76

Published: 2026-09-30T09:10:07Z

Prerelease: no

Release notes:

What's Changed

  • fix(streaming): send S3 temp-credential reads to the bucket's real endpoint by @pwgardipee in https://github.com/Lightning-AI/litData/pull/904
  • chore(deps): update pyarrow requirement from =20.0.0 to >=25.0.1,=0.3 to >=0.5.2 by @dependabot[bot] in https://github.com/Lightning-AI/litData/pull/902
  • chore(deps): update pytest-asyncio requirement from >=1.0.0 to >=1.4.0 by @dependabot[bot] in https://github.com/Lightning-AI/litData/pull/903
  • fix(streaming): refresh credentials against their reported expiry by @pwgardipee in https://github.com/Lightning-AI/litData/pull/906
  • [1/n] Add Code of Conduct and Governance by @deependujha in https://github.com/Lightning-AI/litData/pull/908
  • [3/n] Sync CODEOWNERS with the current maintainer roster by @bhimrazy in https://github.com/Lightning-AI/litData/pull/910
  • [4/n] Simplify contributing guide by @bhimrazy in https://github.com/Lightning-AI/litData/pull/911
  • chore: drop outdated and dead benchmark bot by @bhimrazy in https://github.com/Lightning-AI/litData/pull/912
  • [2/n] specify PyTorch version policy by @deependujha in https://github.com/Lightning-AI/litData/pull/909
  • [4/n] docs: add comparison section between LitData and torchdata in README by @deependujha in https://github.com/Lightning-AI/litData/pull/916
  • [pre-commit.ci] pre-commit suggestions by @pre-commit-ci[bot] in https://github.com/Lightning-AI/litData/pull/807
  • chore(deps): update pytest requirement from ==9.0.* to ==9.1.* by @dependabot[bot] in https://github.com/Lightning-AI/litData/pull/914
  • Fix R2 region defaults and configured client caching by @tchaton in https://github.com/Lightning-AI/litData/pull/920
  • Add built-in temporal array windows to optimize and StreamingDataset by @tchaton in https://github.com/Lightning-AI/litData/pull/922
  • Optimize temporal window reads on POSIX filesystems by @tchaton in https://github.com/Lightning-AI/litData/pull/923
  • fix(streaming): honor index_path for local StreamingDataset dirs by @Yi-111-a in https://github.com/Lightning-AI/litData/pull/917
  • chore(deps): bump the gha-updates group with 2 updates by @dependabot[bot] in https://github.com/Lightning-AI/litData/pull/918
  • fix(processing): prevent output prefix truncation for stable checkpoint filenames by @Yi-111-a in https://github.com/Lightning-AI/litData/pull/926
  • fix(streaming): preserve mounted storage and exact window reads by @tchaton in https://github.com/Lightning-AI/litData/pull/931
  • refactor code for improved readability and consistency by @deependujha in https://github.com/Lightning-AI/litData/pull/930
  • fix(streaming): reset wrapper sample and cycle counters on reset_state_dict by @truecallerabreham in https://github.com/Lightning-AI/litData/pull/928
  • fix(requirements): unpin pyarrow version in extras and test requirements by @deependujha in https://github.com/Lightning-AI/litData/pull/929
  • chore(deps): bump the gha-updates group with 2 updates by @dependabot[bot] in https://github.com/Lightning-AI/litData/pull/924
  • feat(data): add GPU-local CPU and NUMA affinity helpers by @tchaton in https://github.com/Lightning-AI/litData/pull/932
  • feat(io): add opt-in NFS direct reads for Python and native loaders by @tchaton in https://github.com/Lightning-AI/litData/pull/934
  • chore: bump version to 0.2.76 by @tchaton in https://github.com/Lightning-AI/litData/pull/935

New Contributors

  • @Yi-111-a made their first contribution in https://github.com/Lightning-AI/litData/pull/917
  • @truecallerabreham made their first contribution in https://github.com/Lightning-AI/litData/pull/928

Full Changelog: https://github.com/Lightning-AI/litData/compare/v0.2.75...v0.2.76