microsoft/DeepSpeed

Pull Requests Commits

jeffra committed 2 years ago

44d28709

Merge branch 'master' into kv-cache-reset

martincai committed 2 years ago

Verified 0c24ca7c

add support for hjson config files (#2783)

mrwyattii committed 2 years ago

Verified 4079077c

Fix Checkpoint-loading with Meta-tensor (#2781)

RezaYazdaniAminabadi committed 2 years ago

Verified 2c6e8194

Merge branch 'kv-cache-reset' of github.com:microsoft/DeepSpeed into kv-cache-reset

Reza Yazdani committed 2 years ago

1c430eb2

reset cache only if using inferece_cuda module

Reza Yazdani committed 2 years ago

2a2356b0

Merge branch 'master' into kv-cache-reset

RezaYazdaniAminabadi committed 2 years ago

Verified b90d80ca

Fix broken kernel inject bug (#2776)

molly-smith committed 2 years ago

Verified c5b983e9

fix upsample flops compute by skipping unused kargs (#2773)

cli99 committed 2 years ago

Verified b5750b64

some fix in flops_profiler (#2068)

lucasleesw committed 2 years ago

Verified e2a31d80

Fix for diffusers v0.12.0 (#2753)

mrwyattii committed 2 years ago

Verified ef6a958e

Pin minimum `packaging` requirement (#2771)

carmocca committed 2 years ago

Verified 02e95e6a

Refactor/Pydantify monitoring config (#2640)

mrwyattii committed 2 years ago

Verified d923f7c8

Fix hardcoded instances to fp16 in optimizer creation log messages to the correct dtype. (#2743)

loadams committed 2 years ago

Verified 86477538

Add links to new azureML examples (#2756)

cassieesvelt committed 2 years ago

Verified 1db4ade3

Bing/formatting correction (#2764)

xiexbing committed 2 years ago

Verified 8d3b42c2

Add environment variable to make nvcc compilation more verbose (#2759)

cmikeh2 committed 2 years ago

Verified 258d2831

Skip test_bias_gelu unit test if torch < 1.12 (#2754)

lekurile committed 2 years ago

Verified cc3d7cb9

Fix softmax backward (#2709)

RezaYazdaniAminabadi committed 2 years ago

Verified 0b06e0cb

[zero] remove misleading dtype log (#2732)

jeffra committed 2 years ago

Verified a60e31a7

fix a mispelled attribute (#2750)

stas00 committed 2 years ago

Verified 30d3f5df

Fix autotuning so that it records Floating Point Operations per second, not microsecond (#2711)

dashstander committed 2 years ago

Verified d4bfae41

Abstract accelerator (step 3) (#2677)

delock committed 2 years ago

Verified 98cc35b6

[GatheredParameters] fix memory leak (#2665)

stas00 committed 2 years ago

Verified ddd48b36

Merge branch 'master' into kv-cache-reset

RezaYazdaniAminabadi committed 2 years ago

Verified bc00402b

fixing optimizer sanity check (#2742)

jomayeri committed 2 years ago

Verified 4be8df72

Automatic tensor parallelism v2 (#2670)

molly-smith committed 2 years ago

Verified d59b5729

Change zero_grad() argument to match pytorch (#2741)

loadams committed 2 years ago

Verified 34a11688

Inference Refactor (replace_with_policy, model_implementations) (#2554)

awan-10 committed 2 years ago

Verified 867da307

fix typo (#2718)

mrwyattii committed 2 years ago

Verified 8df50a26

Older