microsoft/DeepSpeed

Pull Requests Commits

Use fused addmm and eliminate eye allocation in Gram NS

delock committed 9 days ago

61095611

Fold transpose into matmul in Gram NS for tall matrices

delock committed 9 days ago

a6cf6b69

Fix non-contiguous tensor output from Gram NS for tall matrices

delock committed 9 days ago

e5de42ce

Use accelerator API for dtype selection in Newton-Schulz iterations

delock committed 10 days ago

54930203

fix: correct Gram Newton-Schulz reference URL

delock committed 10 days ago

d17212ef

docs: add ns_method parameter to Muon optimizer documentation

delock committed 10 days ago

e9beb2de

Add Gram Newton-Schulz iteration for Muon optimizer

delock committed 10 days ago

381d8b7a

Merging AutoSP into DeepSpeed (#7860)

neeldani committed 14 days ago

Verified 5efb24ac

Update README with latest news from DeepSpeed (#7931)

PKUWZP committed 14 days ago

Verified 2bae3606

AGENTS.md: Add pre-commit command to existing CI requirements line (#7930)

delock committed 14 days ago

Verified abb88ceb

fix(superoffload) preserve multi-group updates with shared cpu buffers (#7905) (#7906)

xylian86 committed 15 days ago

Verified 729df6ca

Add news item for ASPLOS 2026 Best Paper Award (#7923)

PKUWZP committed 17 days ago

Verified 62c3e6d8

Extending Muon Optimizer Support for ZeRO Stage 3 (#7919)

PKUWZP committed 18 days ago

Verified 956ec6fb

Fix backward compatibility of torch.amp.custom_fwd for PyTorch < 2.4 (#7920)

tohtana committed 18 days ago

Verified 138f20d9

fix: handle non-existent path in is_nfs_path for Triton autotune cache (#7921)

Krishnachaitanyakc committed 18 days ago

Verified f887b982

Add HuggingFace tp_plan support for AutoTP (#7901)

delock committed 19 days ago

Verified a240c4da

Don't detect local GPU if `$DS_IGNORE_CUDA_DETECTION` is set (#7896)

Flamefire committed 19 days ago

Verified 26c954ff

fix: remove unnecessary shell=True in ROCm GPU architecture detection (#7915)

instantraaamen committed 19 days ago

Verified 5ce3abbf

Add Feature Universal Checkpoint for AutoTP (#7908)

nathon-lee committed 20 days ago

Verified f2bb1ec6

Respect `$TRITON_HOME` (#7907)

Flamefire committed 20 days ago

Verified 38bd11a7

Update version (#7903)

loadams committed 30 days ago

Verified 5f7b6870

fix(zero3): use current_stream() instead of default_stream() for grad… (#7898)

michaelroyzen committed 30 days ago

Verified be60451f

Add AGENTS.md and CLAUDE.md with project rules for AI coding agents (#7902)

delock committed 30 days ago

Verified 49a82a0d

fix: Validate fp16.loss_scale is finite and non-negative (#7889)

nathon-lee committed 31 days ago

Verified 63eeb114

Fix Evoformer's multi-arch dispatch root cause (#7881)

tohtana committed 31 days ago

Verified 784cc26e

Fix async_io ops building error on Huawei Ascend NPU (#7894)

huangyifan0610 committed 32 days ago

Verified f88d0f8d

double reduction user-friendly error (#7895)

stas00 committed 33 days ago

Verified b6346bf0

[Bloom] Fix hangs of bloom test (#7890)

k-artem committed 34 days ago

Verified d9a4aadb

Suppress see_memory_usage logs (#7891)

sfc-gh-truwase committed 36 days ago

Verified 285cae30

Fix hook count performance regression from v0.18.5 (#7886)

tohtana committed 38 days ago

Verified 6c59d544

Older