microsoft/DeepSpeed

Pull Requests Commits

Merge branch 'master' into loadams/torch-nightly-debug

loadams committed 1 year ago

Verified 9918386b

fix error type in ccl.py (#4521)

Liangliang-Ma committed 1 year ago

Verified a7358817

Add NPU FusedAdam support (#4343)

CurryRice233 committed 1 year ago

Verified 3e70a887

clear redundant parameters in zero3 bwd hook (#4520)

inkcherry committed 1 year ago

Verified 2cbfb89a

add available memory check to accelerators (#4508)

jeffra committed 1 year ago

Verified 12aedac6

Update README.md (#4518)

samadejacobs committed 1 year ago

Verified 78c518ed

Ulysses: add col-ai evaluation (#4517)

samadejacobs committed 1 year ago

Verified 5bbbf41f

[CCLBackend] update API (#4378)

Liangliang-Ma committed 1 year ago

Verified 4fc181b0

add DeepSpeed4Science white paper (#4502)

conglongli committed 1 year ago

Verified 574fbc0d

Fix scale factor on flops profiler (#4500)

loadams committed 1 year ago

Verified a25a67a0

adding 8bit dequantization kernel for asym fine-grained block quantization in zero-inference (#4450)

stephen-youn committed 1 year ago

Verified 6c86ff39

Revert torch version

loadams committed 1 year ago

a5c1498b

Switch to run on PR to debug

loadams committed 1 year ago

9814448b

Enable running of CI on demand if needed to debug

loadams committed 1 year ago

4b768b35

Update ROCm verison (#4486)

loadams committed 1 year ago

Verified 427253b9

Allow env var for timeout (#4405)

BramVanroy committed 1 year ago

Verified 706a7256

Introduce pydantic_v1 compatibility module for pydantic>=2.0.0 support (#4407)

ringohoffman committed 1 year ago

Verified 604d701e

Remove DS_BUILD_UTILS reference (#4485)

loadams committed 1 year ago

Verified e7acee49

[create-pull-request] automated change (#4484)

mrwyattii committed 1 year ago

Verified 77a28371

fix missing package

mrwyattii committed 1 year ago

Verified e9503fe4

fix bad build command (#4483)

mrwyattii committed 1 year ago

Verified 923f3590

add lm_head and embed_out tensor parallel (#3962)

Yejing-Lai committed 1 year ago

Verified 6763e2de

move torch import (#4468)

mrwyattii committed 1 year ago

Verified 6b634d0e

Move tensors to device if mp is not enabled (#4461)

deepcharm committed 1 year ago

Verified 59d62d61

Fix bug in bfloat16 optimizer related to checkpointing (#4434)

okoge-kaz committed 1 year ago

Verified 7ed952ef

jeffra committed 1 year ago

c4d46795

mrwyattii committed 1 year ago

Verified 26d0dd92

Add release flow (#4467)

loadams committed 1 year ago

Verified 2118c635

pipe engine eval_batch: add option to disable loss broadcast (#4326)

nelyahu committed 1 year ago

Verified f9698c73

feat: add Lion optimizer (#4331)

enneamer committed 1 year ago

Verified 8e64c3b5

Older