microsoft/DeepSpeed

Pull Requests Commits

add option to disable mixed precision

Masahiro Tanaka committed 1 year ago

0a78c7b7

Merge branch 'tohtana/simplify_param_coordinator' of github.com:microsoft/DeepSpeed into tohtana/simplify_param_coordinator

Masahiro Tanaka committed 1 year ago

d24eb643

change test parameter

Masahiro Tanaka committed 1 year ago

acb9efa8

Merge branch 'master' into tohtana/simplify_param_coordinator

tohtana committed 1 year ago

Verified 92a0dd03

Merge branch 'tohtana/simplify_param_coordinator' of github.com:microsoft/DeepSpeed into tohtana/simplify_param_coordinator

Masahiro Tanaka committed 1 year ago

54525e06

test prefetching while switching modes

Masahiro Tanaka committed 1 year ago

04d9dda7

sequence parallel for uneven heads (#6392)

inkcherry committed 1 year ago

Verified 5fb71c0a

Fix dynamo issue (#6527)

oraluben committed 1 year ago

Verified 3d5cf739

Merge branch 'master' into tohtana/simplify_param_coordinator

tohtana committed 1 year ago

Verified 02d044ed

Merge branch 'master' into tohtana/simplify_param_coordinator

tohtana committed 1 year ago

Verified ba3c00c2

remove arg to choose param coordinator

Masahiro Tanaka committed 1 year ago

cb2b25d9

stop tracing for compiler from modifying prefetch info

Masahiro Tanaka committed 1 year ago

c4ac674e

fix initial trace mode

Masahiro Tanaka committed 1 year ago

b3b7bc94

fix init_device_mesh for torch 2.4 (#6614)

Lzhang-hub committed 1 year ago

Verified 6e6563d3

Add attribute check for language_model when replace last linear module (#6650)

Yejing-Lai committed 1 year ago

Verified e06bb518

use one param coordinator for both train/inference modes

Masahiro Tanaka committed 1 year ago

723289f8

Fix expert grad scaling problem with ZeRO optimizer (#6546)

wyooyw committed 1 year ago

Verified b647fb24

Update version.txt after 0.15.3 release (#6652)

loadams committed 1 year ago

Verified bf03f483

[XPU] [DeepNVMe] use same cpu_op_desc_t with cuda (#6645)

Liangliang-Ma committed 1 year ago

Verified a24cdd6b

[XPU] host timer check version from Torch 2.5 to Torch 2.6 (#6633)

YizhouZ committed 1 year ago

Verified 11bbf45a

[XPU] upgrade xpu max1100 CI workflow to pytorch2.3 (#6646)

Liangliang-Ma committed 1 year ago

Verified 40bde528

Fix Memory Leak In AIO (#6630)

jomayeri committed 1 year ago

Verified 6eefc3d0

Use file store for tests (#6632)

tohtana committed 1 year ago

Verified c9fc34a4

Update torch version in workflows (#6631)

tohtana committed 1 year ago

Verified a36db9cc

Add README Pipeline Status for Huawei Ascend NPU (#6588)

xuedinge233 committed 1 year ago

Verified c9899dc1

Lock cache file of HF model list (#6628)

tohtana committed 1 year ago

Verified 1a45bd8e

add option to disable logger while compiling to avoid graph breaks (#6496)

ShellyNR committed 1 year ago

Verified ce468c37

Support safetensors export (#6579)

xu-song committed 1 year ago

Verified bf60fc0c

Add first Step in LR Schedulers (#6597)

jomayeri committed 1 year ago

Verified 85b7469e

Accept btl_tcp_if_include option through launcher_args (#6613)

diskkid committed 1 year ago

Verified 13c16c95

Older