microsoft/DeepSpeed

Pull Requests Commits

Change sot installed cuda version

loadams committed 2 years ago

e7018be3

Fix merge conflicts

loadams committed 2 years ago

2d77a68b

fix for cpu-only tests (#2849)

mrwyattii committed 2 years ago

Verified 6c85fe6c

Fix gpt-Neox rotary embedding implementation (#2782)

RezaYazdaniAminabadi committed 2 years ago

Verified 5b7413a4

Port Reza's INT8-quantization fix to container architecture (#2725)

lekurile committed 2 years ago

Verified fd1449c7

Refactor DS inference API. No longer need replace_method. (#2831)

awan-10 committed 2 years ago

Verified e4b3b610

Fix auto TP for duplicate modules with different gems (#2784)

molly-smith committed 2 years ago

Verified 46784cb5

Fix Pipeline Parallel resize unit test (#2833)

mrwyattii committed 2 years ago

Verified cc1054d9

Fix cpu-only CI hangs (#2825)

mrwyattii committed 2 years ago

Verified 639aa7b3

Fix permissions issue with pip upgrade (#2823)

mrwyattii committed 2 years ago

Verified d3f30199

Fix inference CI device error (#2824)

mrwyattii committed 2 years ago

Verified 93e52897

Handle hanged tests in CI (#2808)

mrwyattii committed 2 years ago

Verified 349f845b

Fix Slurm launcher user args (#2806)

loadams committed 2 years ago

Verified d038dbd2

Add user defined launcher args for PDSH launcher (#2804)

loadams committed 2 years ago

Verified 4af1f76a

Add container load checkpoint error reporting + refactor (#2792)

lekurile committed 2 years ago

Verified 10f3c301

Enable page-locked tensors without CUDA (#2775)

tjruwase committed 2 years ago

Verified c9b08888

remove outdated comment (#2786)

stas00 committed 2 years ago

Verified d323abd8

Fixing broken link to azureml-examples recipes (#2795)

rtanase committed 2 years ago

Verified f376daea

Common location to install libaio-dev (#2779)

tjruwase committed 2 years ago

Verified 7d9fae4d

Container param cleanup + remove qkv_merging (#2780)

lekurile committed 2 years ago

Verified 0a73e6e6

Reset KV-cache at the beginning of text-generation (#2669)

RezaYazdaniAminabadi committed 2 years ago

Verified 9f41ffe4

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

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

Older