Pull Requests microsoft/DeepSpeed

Fix all-gather duplicate params and wrong dtype

#7462 by eternalNight was merged 2025-08-03 00:01

Update version.txt after v0.17.4 release

#7460 by loadams was merged 2025-07-31 21:31

`TiledFusedLogitsLoss` bug fix

#7459 by stas00 was merged 2025-07-31 15:59

Fix invalid f-strings

#7457 by cyyever was merged 2025-08-16 18:22

Update version.txt after 0.17.3 release.

#7455 by loadams was merged 2025-07-28 18:50

Support Muon Optimizer

#7454 by qimcis was closed 2025-08-29 20:30

Fix: UnboundLocalError for variable 'dim' about issue

#7449 by weeknan was merged 2025-07-28 19:18

[ALST] fix typo in the url part2

#7446 by stas00 was merged 2025-07-23 23:31

Remove additional unused tests (human-eval)

#7445 by loadams was merged 2025-07-24 20:16

[ALST] fix typo in the url

#7444 by stas00 was merged 2025-07-23 19:33

Fix: Adapt Llama injection policy for newer transformers versions

#7443 by huanyuqu was merged 2025-07-26 21:27

Remove unused yaml test configurations and update README

#7441 by loadams was merged 2025-07-22 03:56

Use native reduce-scatter for Z1/Z2

#7440 by tohtana was closed 2025-07-22 01:27

adding TiledFusedLogitsLoss

#7437 by stas00 was merged 2025-07-30 18:15

fix issues raised by Coverity scans

#7431 by NirSonnenschein was merged 2025-08-02 16:16

[Ulysses-ALST] add FA3 support

#7430 by stas00 was merged 2025-07-16 15:51

Use past_key_value when provided

#7428 by deepcharm was merged 2025-07-14 20:36

Add getter APIs for TP/PP/DP ranks in DeepSpeedEngine

#7427 by WoosungMyung was merged 2025-08-01 22:23

fix: Propagate `strip_tensor_paddings`

#7426 by saforem2 was merged 2025-07-13 01:22

trying to fix nv-accelerate-v100.yml CI job

#7424 by stas00 was merged 2025-07-11 14:07

TiledMLP + SequenceTiledCompute: improve the bs>1 use-case

#7422 by stas00 was merged 2025-07-16 16:30

Support DeepSpeed offload and reload states with ZeRO1 and ZeRO2

#7421 by LYMDLUT was merged 2025-08-20 22:03

[BUGFIX] Reset `bucket.elements` after reduction in ZeRO Stage 3

#7418 by rahul713rk was merged 2025-07-08 02:32

Update version.txt after v0.17.2 release.

#7417 by loadams was merged 2025-07-07 19:10

Align missing argument in AllReduceCoalescedHandle

#7414 by deepcharm was merged 2025-07-08 21:10

[TiledMLP]: fix for bs>1

#7412 by stas00 was merged 2025-07-07 18:34

fix: engine initializes optimizer attributes at the beginning

#7410 by HollowMan6 was merged 2025-07-07 02:24

Improvements to Communication Logger

#7404 by alexk101 was merged 2025-07-09 17:35

fix: DeepCompile for torch 2.8

#7402 by HollowMan6 was merged 2025-06-30 18:19

fix: Add `csrc/compile` to include paths for DeepCompile builder

#7401 by HollowMan6 was merged 2025-06-30 17:55