microsoft/DeepSpeed

Pull Requests Commits

loadams committed 1 year ago

504e1b64

Increase CI coverage for Gaudi2 accelerator. (#5358)

vshekhawat-hlab committed 1 year ago

Verified ca4bc66b

Update pytest and transformers with fixes for pytest>= 8.0.0 (#5164)

loadams committed 1 year ago

Verified 72bd275b

Support MoE for pipeline models (#5338)

mosheisland committed 1 year ago

Verified 08e0733e

Auto convert moe param groups (#5354)

jeffra committed 1 year ago

Verified 42a8eaa7

CPU SHM based inference_all_reduce improve (#5320)

delock committed 1 year ago

Verified 731fd682

FP [6,8,12] quantizer op (#5336)

jeffra committed 1 year ago

Verified 3fbd01cc

Update workflows that use cu116 to cu117 (#5361)

loadams committed 1 year ago

Verified 4621ba4c

Ensure capacity does not exceed number of tokens (#5353)

jeffra committed 1 year ago

Verified 5fbc3eee

logger update with torch master changes (#5346)

rogerxfeng8 committed 1 year ago

Verified 6a4b96a1

compute global norm on device (#5125)

BacharL committed 1 year ago

Verified 9f0e2136

Update container version for Gaudi2 CI (#5360)

raza-sikander committed 1 year ago

Verified 548d37b1

[fix] fix typo s/simultanenously /simultaneously (#5359)

digger-yu committed 1 year ago

Verified d86a68c3

Add `distributed_port` for `deepspeed.initialize` (#5260)

LZHgrla committed 1 year ago

Verified dc3554f8

Fix sort of zero checkpoint files (#5342)

tohtana committed 1 year ago

Verified c946a342

BF16 optimizer: Clear lp grads after updating hp grads in hook (#5328)

YangQun1 committed 1 year ago

Verified 40009eb1

resolve KeyError: 'PDSH_SSH_ARGS_APPEND' (#5318)

Lzhang-hub committed 1 year ago

Verified cc897ecf

Make NFS warning print only once (#5345)

HeyangQin committed 1 year ago

Verified b5e2045c

Revert "Rename test group from amd-test to unit-tests to make rrequired when enabled and running on PRs"

loadams committed 1 year ago

1bc15efe

Rename test group from amd-test to unit-tests to make rrequired when enabled and running on PRs

loadams committed 1 year ago

378440d9

Correct typo in checking on bf16 unit test support (#5317)

loadams committed 1 year ago

Verified 014576a8

Add NFS path check for default deepspeed triton cache directory (#5323)

HeyangQin committed 1 year ago

Verified ffb53c25

Increase coverage for HPU (#5324)

loadams committed 1 year ago

Verified 4c51dce8

Improve universal checkpoint (#5289)

tohtana committed 1 year ago

Verified c56a4b9e

Added Gaudi2 CI tests. (#5275)

vshekhawat-hlab committed 1 year ago

Verified 330d36bb

Docs typos fix and grammar suggestions (#5322)

Gr0g0 committed 1 year ago

Verified cea5ea1e

Modified regular expression (#5306)

igeni committed 1 year ago

Verified a32789b0

fix comms dtype (#5297)

mayank31398 committed 1 year ago

Verified ed8aed57

Enable mixtral 8x7b autotp (#5257)

Yejing-Lai committed 1 year ago

Verified 8d98e171

support bf16_optimizer moe expert parallel training and moe EP grad_scale/grad_norm fix (#5259)

inkcherry committed 1 year ago

Verified e5dd5501

Older