microsoft/DeepSpeed

Pull Requests Commits

Only run pre-commit on the changes (#7106)

hwchen2017 committed 282 days ago

Verified e4c7931a

Avoid graph breaks in torch.compile caused by inner classes in the backward hooks (#7062)

deepcharm committed 283 days ago

Verified 776822fe

Avoid graph breaks by disabling sourceless calls in instrument_w_nvtx (#7081)

deepcharm committed 283 days ago

Verified a88f56a0

Use new dlpack api; Formatting fixes (#7101)

tjruwase committed 283 days ago

Verified b4177e45

Remove workflows for very old torch versions (#7090)

loadams committed 287 days ago

Verified 02bbf501

Update parallelism for nv-torch-latest/nightly tests due to more GPUs/runner (#7086)

loadams committed 287 days ago

Verified f2ed2531

Update README.md with ICS '23 MoE paper link (#7087)

siddharth9820 committed 287 days ago

Verified 5320d4c9

Pin transformers version on tests that use latest. (#7085)

loadams committed 287 days ago

Verified f8d34295

Improve inference tutorial docs (#7083)

loadams committed 288 days ago

Verified c07b635c

Add DeepseekV3 AutoTP. (#7045)

Yejing-Lai committed 288 days ago

Verified f0401ade

Revert "Handle special case of libuv for Windows (#7064)" (#7076)

loadams committed 290 days ago

Verified 729dfafc

Replace calls to `python setup.py sdist` with `python -m build --sdist` (#7069)

loadams committed 290 days ago

Verified 1d30b58c

Fix meta load tensor imcompatible issue (#7073)

Yejing-Lai committed 290 days ago

Verified 4b7e2c90

config torch to avoid graph breaks caused by logger (#6999)

ShellyNR committed 290 days ago

Verified e1903f0d

Fix TOCTOU issues, switch to fstat (#7067)

loadams committed 291 days ago

Verified 9d820e4c

Bug Fix for offload_states API (#7050)

U-rara committed 293 days ago

Verified 38327e07

Update README with info on newest accelerator (#7065)

loadams committed 293 days ago

Verified 9f20148a

Handle special case of libuv for Windows (#7064)

loadams committed 294 days ago

Verified 8577bd24

Tecorigin sdaa accelerator (#6903)

siqi654321 committed 294 days ago

Verified cb20d449

fix an outdated doc wrt CUDA_VISIBLE_DEVICES (#7058)

stas00 committed 294 days ago

Verified 461d641f

Update version.txt after 0.16.4 release (#7063)

loadams committed 294 days ago

Verified fa8967e2

Fix, bf16 optimizer remove dup loop (#7054)

wukong1992 committed 294 days ago

Verified e2dc3eeb

add autoTP training zero2 tests (#7049)

inkcherry committed 295 days ago

Verified d98204b9

Rename aio_thread_count to intra_op_parallelism (#7056)

tjruwase committed 295 days ago

Verified c9da4892

nv-ds-chat breaks with latest transformers (#7052)

loadams committed 295 days ago

Verified 33dd2e21

Update setup.py handling of ROCm cupy (#7051)

loadams committed 296 days ago

Verified 7288e619

Update CUDA compute capability to support Blackwell (#7047)

hwchen2017 committed 296 days ago

Verified 2735cf41

Control trace cache warnings (#7039)

tjruwase committed 296 days ago

Verified ee3f19bf

AIO on ROCM (#7023)

jomayeri committed 300 days ago

Verified 14b3cce4

Add chinese blog for deepspeed windows, and fix format (#7035)

hwchen2017 committed 301 days ago

Verified e6376777

Newer Older