microsoft/DeepSpeed

Pull Requests Commits

Merge branch 'master' into styoun/zero-inf-8bit-q

tjruwase committed 2 years ago

Verified c9704348

Update ROCm verison (#4486)

loadams committed 2 years ago

Verified 427253b9

Allow env var for timeout (#4405)

BramVanroy committed 2 years ago

Verified 706a7256

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

Matthew Hoffman committed 2 years ago

Verified 604d701e

Remove DS_BUILD_UTILS reference (#4485)

loadams committed 2 years ago

Verified e7acee49

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

mrwyattii committed 2 years ago

Verified 77a28371

fix missing package

mrwyattii committed 2 years ago

Verified e9503fe4

Merge branch 'styoun/zero-inf-8bit-q' of https://github.com/microsoft/DeepSpeed into styoun/zero-inf-8bit-q

styoun committed 2 years ago

7c7310dc

styoun committed 2 years ago

68c258c3

fix bad build command (#4483)

mrwyattii committed 2 years ago

Verified 923f3590

add lm_head and embed_out tensor parallel (#3962)

Yejing-Lai committed 2 years ago

Verified 6763e2de

move torch import (#4468)

mrwyattii committed 2 years ago

Verified 6b634d0e

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

deepcharm committed 2 years ago

Verified 59d62d61

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

okoge-kaz committed 2 years ago

Verified 7ed952ef

Merge branch 'master' into styoun/zero-inf-8bit-q

tjruwase committed 2 years ago

Verified 571ab839

increased coverage of QuantLinear test

styoun committed 2 years ago

0b216378

fix after the pr comments

styoun committed 2 years ago

1ea4bf45

jeffra committed 2 years ago

c4d46795

mrwyattii committed 2 years ago

Verified 26d0dd92

Add release flow (#4467)

loadams committed 2 years ago

Verified 2118c635

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

nelyahu committed 2 years ago

Verified f9698c73

feat: add Lion optimizer (#4331)

enneamer committed 2 years ago

Verified 8e64c3b5

fix lm head overriden issue, move it from checkpoint in-loop loading to out loop (#4206)

sywangyi committed 2 years ago

Verified d72edb3b

CI fix for torch 2.1 release (#4452)

mrwyattii committed 2 years ago

Verified 4294ea17

styoun committed 2 years ago

359c3e58

"rename test_int4_quantization.py to test_intX_quantization.py"

styoun committed 2 years ago

b9896cd3

rename test_int4_quantization.py to test_intX_quantization.py

styoun committed 2 years ago

2d341405

Merge branch 'styoun/zero-inf-8bit-q' of https://github.com/microsoft/DeepSpeed into styoun/zero-inf-8bit-q

styoun committed 2 years ago

90858e7c

rename quantize_int4.cu to quantize_intX.cu

styoun committed 2 years ago

e0ad1673

DeepSpeed-VisualChat Chinese blog (#4458)

conglongli committed 2 years ago

Verified 2c220d65

Older