Pull Requests deepspeedai/DeepSpeed

Improve error message and reduce validation in autocast test

#7547 by tohtana was merged 2025-09-05 07:04

Fix zenflow_torch_adam.py

#7544 by stas00 was merged 2025-09-03 18:14

Relax restrictions of torch.autocast integration

#7543 by tohtana was merged 2025-09-03 19:15

avoid setting device_id to `init_process_group`

#7542 by kaixuanliu was merged 2025-09-05 06:06

[Deepspeed]Fix Muon optimizer import error

#7539 by pengdurice was closed 2025-09-02 18:23

[doc] fixing moe tutorial

#7538 by stas00 was merged 2025-09-02 20:53

undo the revert

#7536 by stas00 was merged 2025-09-02 18:24

Fix scaling and allgather with `torch.autocast`

#7534 by tohtana was merged 2025-09-03 01:22

fixed DeepSpeedCPULion with ZeRO-Offload bug

#7531 by qibin0506 was merged 2025-09-02 21:40

docs typo: `lrrt.md`, reference to `cycle_min_lr` should be `cycle_max_lr`

#7530 by jakehemmerle was merged 2025-09-02 21:17

fix typo s/1014 /1024

#7528 by digger-yu was merged 2025-09-01 01:12

[logging] less startup noise

#7526 by stas00 was merged 2025-09-02 19:14

ZeRO3: Improve mismatch detection

#7525 by sfc-gh-truwase was merged 2025-08-31 21:57

fix get_cuda_compile_flag

#7521 by mingjielu was merged 2025-09-04 16:34

Update README with ZenFlow release blog featured by PyTorch.

#7520 by Antlera was merged 2025-08-28 17:28

Add riscv64 cpu support in deepspeed_shm_comm op

#7519 by heyujiao99 was merged 2025-08-29 15:41

Update TSC Committers

#7517 by PKUWZP was merged 2025-08-26 11:24

Enable non-ZeRO mode

#7515 by sfc-gh-truwase was merged 2025-08-27 18:07

Enabling Muon Optimizer in DeepSpeed

#7509 by PKUWZP was merged 2025-08-27 01:34

Fix assert when 'pp_int' object has no attribute 'custom_print_str'

#7507 by aeeeeeep was merged 2025-08-25 14:57

Autotune ZenFlow affinity

#7506 by delock was merged 2025-09-04 11:10

fix `_disable_dynamo_if_unsupported` not found

#7505 by ZhiyuanChen was closed 2025-09-03 18:56

CI funding shout out to modal.com

#7503 by stas00 was merged 2025-08-21 17:03

Update version.txt after 0.17.5 release

#7502 by loadams was merged 2025-08-20 21:41

Reduce performance impact of compiler.enable decorator

#7498 by deepcharm was merged 2025-08-18 22:04

Add index to HPU devices

#7497 by deepcharm was merged 2025-08-19 00:30

Fix DeepCompile for PyTorch v2.8

#7496 by tohtana was merged 2025-08-18 16:13

fix xpu device_id AttributeError issue

#7488 by yao-matrix was merged 2025-08-15 16:40

Enable forked PRs

#7486 by sfc-gh-truwase was merged 2025-08-14 21:43

Enable forked PRs to run on Modal

#7485 by sfc-gh-truwase was closed 2025-08-12 13:34