microsoft/DeepSpeed

Pull Requests Commits

Update README.md

HeyangQin committed 1 year ago

03e797eb

Merge branch 'master' into HeyangQin/ucp_blog_chinese

HeyangQin committed 1 year ago

07f2d3d8

DeepSpeed Universal Checkpointing: Blog and Tutorial (#5711)

samadejacobs committed 1 year ago

Verified 121efdbd

update xpu fusedadam opbuilder for pytorch 2.3 (#5702)

baodii committed 1 year ago

Verified e3922967

Update XPU docker version (#5712)

loadams committed 1 year ago

Verified df58a784

Update README.md

HeyangQin committed 1 year ago

9db1d4b1

HeyangQin committed 1 year ago

673c43ae

Update README.md

samadejacobs committed 1 year ago

Verified af1e01c4

Update index.md

samadejacobs committed 1 year ago

Verified 1f74a48d

Update README.md

samadejacobs committed 1 year ago

Verified e77a7881

Update README.md

samadejacobs committed 1 year ago

Verified f2b43aff

add Japanese translation of UCP blog

Masahiro Tanaka committed 1 year ago

be756a74

Update index page

samadejacobs committed 1 year ago

5da9fe27

Update index page

samadejacobs committed 1 year ago

aa829f14

Update index page

samadejacobs committed 1 year ago

923a46b1

Add tutorial and references to Chinese and Japanese Twitter handles.

samadejacobs committed 1 year ago

b5405a00

samadejacobs committed 1 year ago

7b6c9b6d

DeepSpeed Universal Checkpointing blog

samadejacobs committed 1 year ago

a2a4175f

Add additional paths to trigger xpu tests (#5707)

loadams committed 1 year ago

Verified aecfec7f

Change source of CPUAdam for xpu accelerator (#5703)

Liangliang-Ma committed 1 year ago

Verified 4b8a4a07

Add an argument to enable the injection of missing state during the conversion of universal checkpoints (#5608)

xylian86 committed 1 year ago

Verified f0e3f01d

Disable nvtx decorator to avoid graph break (#5697)

tohtana committed 1 year ago

Verified b421e8c8

ENV var added for recaching in INF Unit tests (#5688)

raza-sikander committed 1 year ago

Verified e9ffe028

inference unit test injectionPolicy split world_size to multiple tests (#5687)

oelayan7 committed 1 year ago

Verified 4f8531bc

Universal checkpoint for zero stage 3 (#5475)

xylian86 committed 1 year ago

Verified d2b1d7fc

[CPU] add fp16 support to shm inference_all_reduce (#5669)

delock committed 1 year ago

Verified 19da95f7

Fix memory leak from _hp_mapping (#5643)

chiragjn committed 1 year ago

Verified 88b2ef71

Bug fix for the "Link bit16 and fp32 parameters in partition" (#5681)

U-rara committed 1 year ago

Verified 224a05c3

Fixed Windows inference build. (#5609)

costin-eseanu committed 1 year ago

Verified b3767d01

Update version.txt after 0.14.4 release (#5694)

mrwyattii committed 1 year ago

Verified d89e8cdf

Older