microsoft/DeepSpeed

Pull Requests Commits

Update new_quantize.cu

GuanhuaWang committed 2 years ago

Verified 1fa4a22d

Update pt_binding.cpp

GuanhuaWang committed 2 years ago

Verified e238121e

Update custom_cuda_layers.h

GuanhuaWang committed 2 years ago

Verified b6f350e3

Update quantizer.py

GuanhuaWang committed 2 years ago

Verified bc8b8478

Update pt_binding.cpp

GuanhuaWang committed 2 years ago

Verified 3d72a940

Update new_quantize.cu

GuanhuaWang committed 2 years ago

3e8a28ca

GuanhuaWang committed 2 years ago

Verified f4996ed6

finish quant dequant test

GuanhuaWang committed 2 years ago

92b6390f

Add tensor parallel inference unit tests (#2232)

mrwyattii committed 2 years ago

Verified a691ec60

Update relu.cu with mem_access_utils (#2306)

molly-smith committed 2 years ago

Verified d0dfe38d

Unit test for bias add kernel (#2298)

mrwyattii committed 2 years ago

Verified b2d550ab

Fix for Zero3 when MP>1 and at least one batch param undefined (#2289)

Quentin-Anthony committed 2 years ago

Verified 4524b8db

fix for pytest picking up local deepspeed dir instead of installed deepspeed (#2299)

mrwyattii committed 2 years ago

Verified 7e085b62

fused bias relu unittest (#2297)

molly-smith committed 2 years ago

Verified 5102c4ab

Offload all gradients to nvme (#2282)

tjruwase committed 2 years ago

Verified c8641032

[ds-inference] fix progress bar (#2286)

stas00 committed 2 years ago

Verified b146aa35

Refactor universal checkpointing and tensor fragments (#2253)

tjruwase committed 2 years ago

Verified 53182531

Fp32 accuracy bug fix (#2285)

RezaYazdaniAminabadi committed 2 years ago

Verified 47e030f5

Memory Access Utility (#2276)

cmikeh2 committed 2 years ago

Verified c84bca37

Add to codeowners file (#2279)

cmikeh2 committed 2 years ago

Verified 5505e247

add a new unit test for cuda ops (#2278)

awan-10 committed 2 years ago

Verified 91eba15a

Ds-inference Int8 support through ZeroQuant technology (#2217)

RezaYazdaniAminabadi committed 2 years ago

Verified afdc7287

prevent torch ext folder mkdir at tmp (#2274)

jeffra committed 2 years ago

Verified 088212a7

Trajepl: nebula load fix (#2182)

trajepl committed 2 years ago

Verified aca34a9b

Ds inference/fix mp2 (#2270)

RezaYazdaniAminabadi committed 2 years ago

Verified d154cc0f

Fix OrderedDict import for python3.6 (#2267)

Dipet committed 2 years ago

Verified 4671cce5

Update replace_module.py, test-gptj.py related fix (#2269)

molly-smith committed 2 years ago

Verified a7ee688a

Add blob storage to CI runners (#2260)

mrwyattii committed 2 years ago

Verified 55b7b9e0

jeffra committed 2 years ago

220e5088

fix #2240: wrong time unit in flops_profiler (#2241)

yzs981130 committed 2 years ago

Verified 80f94c10

Older