MORPH
SIGN IN SIGN UP
deepspeedai / DeepSpeed UNCLAIMED

DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.

41918 0 0 Python

ACTIVITY

week month year
COMMITS
20
in the last month
CONTRIBUTORS
13
active
STARS
41918
total
FORKS
0
total
TOP CONTRIBUTORS
M
Masahiro Tanaka
4 commits
Z
Zhipeng Wang
2 commits
M
Ma, Guokai
2 commits
A
Alexander Grund
2 commits
N
nathon
2 commits
K
Krishna Chaitanya
1 commit
I
instantraaamen
1 commit
L
Logan Adams
1 commit
M
Michael Royzen
1 commit
H
Huang Yifan
1 commit
RECENT COMMITS
Z
Add news item for ASPLOS 2026 Best Paper Award (#7923)
Zhipeng Wang
62c3e6d
Z
Extending Muon Optimizer Support for ZeRO Stage 3 (#7919)
Zhipeng Wang
956ec6f
M
Fix backward compatibility of torch.amp.custom_fwd for PyTorch < 2.4 (#7920)
Masahiro Tanaka
138f20d
K
fix: handle non-existent path in is_nfs_path for Triton autotune cache (#7921)
Krishna Chaitanya
f887b98
M
Add HuggingFace tp_plan support for AutoTP (#7901)
Ma, Guokai
a240c4d
A
Don't detect local GPU if `$DS_IGNORE_CUDA_DETECTION` is set (#7896)
Alexander Grund
26c954f
I
fix: remove unnecessary shell=True in ROCm GPU architecture detection (#7915)
instantraaamen
5ce3abb
N
Add Feature Universal Checkpoint for AutoTP (#7908)
nathon
f2bb1ec
A
Respect `$TRITON_HOME` (#7907)
Alexander Grund
38bd11a
L
Update version (#7903)
Logan Adams
5f7b687
M
fix(zero3): use current_stream() instead of default_stream() for grad… (#7898)
Michael Royzen
be60451
M
Add AGENTS.md and CLAUDE.md with project rules for AI coding agents (#7902)
Ma, Guokai
49a82a0
N
fix: Validate fp16.loss_scale is finite and non-negative (#7889)
nathon
63eeb11
M
Fix Evoformer's multi-arch dispatch root cause (#7881)
Masahiro Tanaka
784cc26
H
Fix async_io ops building error on Huawei Ascend NPU (#7894)
Huang Yifan
f88d0f8
S
double reduction user-friendly error (#7895)
Stas Bekman
b6346bf
A
[Bloom] Fix hangs of bloom test (#7890)
Artem Kuzmitckii
d9a4aad
O
Suppress see_memory_usage logs (#7891)
Olatunji Ruwase
285cae3
M
Fix hook count performance regression from v0.18.5 (#7886)
Masahiro Tanaka
6c59d54
M
Add document section explaining autocast nesting (#7883)
Masahiro Tanaka
4dba1e2