Download Latest Version
v0.18.9 Patch Release source code.tar.gz (216.3 MB)
Get an email when there's a new version of DeepSpeed
| Name | Modified | Size | Downloads / Week |
|---|---|---|---|
| Parent folder | |||
| README.md | 2026-02-12 | 1.5 kB | |
| v0.18.6 Patch Release source code.tar.gz | 2026-02-12 | 216.2 MB | |
| v0.18.6 Patch Release source code.zip | 2026-02-12 | 217.4 MB | |
| Totals: 3 Items | 433.6 MB | 1 | |
What's Changed
- Update version.txt to 0.18.6 after latest release by @loadams in https://github.com/deepspeedai/DeepSpeed/pull/7826
- Fix leaf module race condition by @tohtana in https://github.com/deepspeedai/DeepSpeed/pull/7825
- Skip sequence parallel operations during eval by @jp1924 in https://github.com/deepspeedai/DeepSpeed/pull/7821
- Support custom partitioning patterns for AutoTP by @tohtana in https://github.com/deepspeedai/DeepSpeed/pull/7806
- Fix gradient is ready with z2 by @sfc-gh-truwase in https://github.com/deepspeedai/DeepSpeed/pull/7829
- Fix AutoTP custom patterns: respect use_default_specs by @tohtana in https://github.com/deepspeedai/DeepSpeed/pull/7827
- Support new python 3.14 annotation handling by @sdvillal in https://github.com/deepspeedai/DeepSpeed/pull/7831
- fix: replace deprecated fractions.gcd with math.gcd by @Mr-Neutr0n in https://github.com/deepspeedai/DeepSpeed/pull/7845
- Fix bf16 gradient norm divergence with ZeRO stage 0 by @tohtana in https://github.com/deepspeedai/DeepSpeed/pull/7839
- Replace torch.jit.script with torch.compile (#7835) by @tohtana in https://github.com/deepspeedai/DeepSpeed/pull/7840
New Contributors
- @jp1924 made their first contribution in https://github.com/deepspeedai/DeepSpeed/pull/7821
- @Mr-Neutr0n made their first contribution in https://github.com/deepspeedai/DeepSpeed/pull/7845
Full Changelog: https://github.com/deepspeedai/DeepSpeed/compare/v0.18.5...v0.18.6
