microsoft/DeepSpeed

Pull Requests Commits

Merge branch 'master' into loadams/clear-cache

loadams committed 1 year ago

Verified d33b0821

Update other workflows to run on Ubuntu 22.04 (#5798)

loadams committed 1 year ago

Verified ffd0a0e3

Unpin transformers version (#5650)

loadams committed 1 year ago

Verified e661ecb3

move is_checkpointable call reducing torch.compile Graph breaks (#5759)

NirSonnenschein committed 1 year ago

Verified 6d0dbf86

Update torch version in cpu-torch-latest and nv-torch-latest-v100 tests to 2.4 (#5797)

loadams committed 1 year ago

Verified bf696ab6

Fix accuracy error of NPUFusedAdam (#5777)

penn513 committed 1 year ago

Verified 5a100f6b

Add new autotp supported model in doc (#5785)

Yejing-Lai committed 1 year ago

Verified acdf1367

Add chatglm2 & chatglm3 autotp (#5540)

Yejing-Lai committed 1 year ago

Verified 0d3bb77b

fix: quantization with DeepSpeed HE (#5624)

Atry committed 1 year ago

Verified 9fa4c424

[INF] Add Qwen2RMSNorm to loaded layers in auto_tp (#5786)

oelayan7 committed 1 year ago

Verified 830d0c0a

Update Ubuntu version for running python tests (#5783)

loadams committed 1 year ago

Verified 85c66fd7

reduce all-to-all communication volume when both expert and non-expert are tensor-parallel (#5626)

taozhiwei committed 1 year ago

Verified f5d6c631

loadams committed 1 year ago

1479d1e9

Fixes for latest Huggingface_hub changes on modelId -> id (#5789)

loadams committed 1 year ago

Verified 213e2d97

Misplaced global variable `warned` (#5725)

anferico committed 1 year ago

Verified 879c6cd0

Add support for Microsoft Phi-3 model to DeepSpeed-FastGen (#5559)

adk9 committed 1 year ago

Verified 6a163e03

Update the list of supported models in the Chinese README of fastgen (#5773)

beep-bebop committed 1 year ago

Verified 78c6c449

Launcher mode with SSH bypass (#5728)

dogacancolak-kensho committed 1 year ago

Verified acbaca32

[bugfix] promote state in bf16_optimizer (#5767)

billishyahao committed 1 year ago

Verified 98272d14

Added wrappers for hpu tensors based on dtype (#5771)

deepcharm committed 1 year ago

Verified 61e07786

[CPU] Allow deepspeed.comm.inference_all_reduce in torch.compile graph (#5604)

delock committed 1 year ago

Verified ec6cbb3c

Fix phi3 mini 128k load error (#5765)

Yejing-Lai committed 1 year ago

Verified a07a3c5d

Remove duplicated variable (#5727)

xu-song committed 1 year ago

Verified 0af9ac31

Fix memory leak for pipelined optimizer swapper (#5700)

mauryaavinash95 committed 1 year ago

Verified db5a875b

Unit Test: Add error handling for rate limit exceeded in model list (#5715)

HeyangQin committed 1 year ago

Verified 83aa1843

Add Windows scripts (deepspeed, ds_report). (#5699)

costin-eseanu committed 1 year ago

Verified 74f3dcab

[INF] Enable torch compile for inference (#5612)

oelayan7 committed 1 year ago

Verified 7b1ea225

Update checkout action for nv-human-eval workflow (#5757)

loadams committed 1 year ago

Verified 2105976e

fix the missing argument in test and typo (#5730)

xylian86 committed 1 year ago

Verified 774b8977

Update node16 check on self-hosted runners and remove python 3.6 (#5756)

loadams committed 1 year ago

Verified 84118165

Older