Download Latest Version
v0.17.4 Patch Release source code.tar.gz (199.7 MB)
Get an email when there's a new version of DeepSeed
| Name | Modified | Size | Downloads / Week |
|---|---|---|---|
| Parent folder | |||
| README.md | 2025-07-26 | 2.3 kB | |
| v0.17.3 Patch Release source code.tar.gz | 2025-07-26 | 199.7 MB | |
| v0.17.3 Patch Release source code.zip | 2025-07-26 | 200.8 MB | |
| Totals: 3 Items | 400.5 MB | 1 | |
What's Changed
- [TiledMLP]: fix for bs>1 by @stas00 in https://github.com/deepspeedai/DeepSpeed/pull/7412
- Update version.txt after v0.17.2 release. by @loadams in https://github.com/deepspeedai/DeepSpeed/pull/7417
- Enable torch version dependent compilation of record_module and iter_params by @deepcharm in https://github.com/deepspeedai/DeepSpeed/pull/7362
- [BUGFIX] Reset
bucket.elementsafter reduction in ZeRO Stage 3 by @rahul713rk in https://github.com/deepspeedai/DeepSpeed/pull/7418 - Align missing argument in AllReduceCoalescedHandle by @deepcharm in https://github.com/deepspeedai/DeepSpeed/pull/7414
- Improvements to Communication Logger by @alexk101 in https://github.com/deepspeedai/DeepSpeed/pull/7404
- trying to fix nv-accelerate-v100.yml CI job by @stas00 in https://github.com/deepspeedai/DeepSpeed/pull/7424
- fix: Propagate
strip_tensor_paddingsby @saforem2 in https://github.com/deepspeedai/DeepSpeed/pull/7426 - Use past_key_value when provided by @deepcharm in https://github.com/deepspeedai/DeepSpeed/pull/7428
- set
device_idin torch'sinit_process_groupby @stas00 in https://github.com/deepspeedai/DeepSpeed/pull/7266 - [Ulysses-ALST] add FA3 support by @stas00 in https://github.com/deepspeedai/DeepSpeed/pull/7430
- TiledMLP + SequenceTiledCompute: improve the bs>1 use-case by @stas00 in https://github.com/deepspeedai/DeepSpeed/pull/7422
- Remove unused yaml test configurations and update README by @loadams in https://github.com/deepspeedai/DeepSpeed/pull/7441
- [ALST] fix typo in the url by @stas00 in https://github.com/deepspeedai/DeepSpeed/pull/7444
- [ALST] fix typo in the url part2 by @stas00 in https://github.com/deepspeedai/DeepSpeed/pull/7446
- Remove additional unused tests (human-eval) by @loadams in https://github.com/deepspeedai/DeepSpeed/pull/7445
- Fix: Adapt Llama injection policy for newer transformers versions by @huanyuqu in https://github.com/deepspeedai/DeepSpeed/pull/7443
New Contributors
- @rahul713rk made their first contribution in https://github.com/deepspeedai/DeepSpeed/pull/7418
- @huanyuqu made their first contribution in https://github.com/deepspeedai/DeepSpeed/pull/7443
Full Changelog: https://github.com/deepspeedai/DeepSpeed/compare/v0.17.2...v0.17.3