microsoft/DeepSpeed

Pull Requests Commits

Merge branch 'master' into olruwase/cuda_optional_ops

tjruwase committed 2 years ago

Verified 813dbba2

[Bug Fixed] use torch.cuda.is_available() (#2661)

wkcn committed 2 years ago

Verified 323c266c

Merge branch 'master' into olruwase/cuda_optional_ops

tjruwase committed 2 years ago

Verified fcf8c333

Remove unnecessary device synchronization for stage 2 (#2500)

li-yi-dong committed 2 years ago

Verified 97deaaec

Merge branch 'master' into olruwase/cuda_optional_ops

tjruwase committed 2 years ago

Verified f654f6a0

Abstract accelerator (step 2) (#2560)

delock committed 2 years ago

Verified 9548d48f

Fix Opt injection (#2541)

RezaYazdaniAminabadi committed 2 years ago

Verified 95d9a1b6

[launcher] fail gracefully if hostname -i doesn't work as expected (#2631)

jeffra committed 2 years ago

Verified a091bc22

[doc] fix `min_loss_scale` default (#2660)

stas00 committed 2 years ago

Verified f2ea7a38

tweaks to ds-attn, distilbert policy, and mup (#2649)

jeffra committed 2 years ago

Verified d9b788d7

tjruwase committed 2 years ago

58d41068

fix assertion error in zero stage 3 (#2647)

GuanhuaWang committed 2 years ago

Verified 6375cb3f

Fix issue w. bloom when changing tp size (#2645)

jeffra committed 2 years ago

Verified e0aa84c5

[inference] ds-attention refactor w.r.t. ops (#2623)

jeffra committed 2 years ago

Verified bb68c526

[zero-3] Handle forward parameter return correctly in nested cases (#2642)

samyam committed 2 years ago

Verified a298a43a

tjruwase committed 2 years ago

3583f97f

Address PR feedback

tjruwase committed 2 years ago

e1f73725

Update .github/workflows/nv-torch-latest-cpu.yml

tjruwase committed 2 years ago

Verified 7c14844d

Update .github/workflows/nv-torch-latest-cpu.yml

tjruwase committed 2 years ago

Verified 1ebb1304

Merge branch 'master' into olruwase/cuda_optional_ops

jeffra committed 2 years ago

Verified 7fe7d500

Fix typo in autotuner.py (#2639)

eltociear committed 2 years ago

Verified 6273dffc

Merge branch 'master' into olruwase/cuda_optional_ops

tjruwase committed 2 years ago

Verified 86d72a4e

Merge branch 'olruwase/cuda_optional_ops' of github.com:microsoft/DeepSpeed into olruwase/cuda_optional_ops

tjruwase committed 2 years ago

8da0c267

Fix cpu_adam unit test

tjruwase committed 2 years ago

8ca082d1

add enable_each_rank_log to deepspeed/launcher/runner.py (#2571)

inkcherry committed 2 years ago

Verified 11f5daba

Merge branch 'master' into olruwase/cuda_optional_ops

tjruwase committed 2 years ago

Verified 5ed6ef1b

Unit tests setup own venv (#2628)

mrwyattii committed 2 years ago

Verified 6fff50f5

Remove scattered pytest.skip

tjruwase committed 2 years ago

fea4d85f

Merge branch 'olruwase/cuda_optional_ops' of github.com:microsoft/DeepSpeed into olruwase/cuda_optional_ops

tjruwase committed 2 years ago

613ca7e6

tjruwase committed 2 years ago

c3a39997

Newer Older