microsoft/DeepSpeed

Pull Requests Commits

Merge branch 'master' into tohtana/clean_all_param_coordinators

tohtana committed 213 days ago

71457520

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

wyooyw committed 214 days ago

Verified b647fb24

Update version.txt after 0.15.3 release (#6652)

loadams committed 214 days ago

Verified bf03f483

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

Liangliang-Ma committed 214 days ago

Verified a24cdd6b

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

YizhouZ committed 215 days ago

Verified 11bbf45a

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

Liangliang-Ma committed 216 days ago

Verified 40bde528

Fix Memory Leak In AIO (#6630)

jomayeri committed 219 days ago

Verified 6eefc3d0

Use file store for tests (#6632)

tohtana committed 219 days ago

Verified c9fc34a4

Update torch version in workflows (#6631)

tohtana committed 219 days ago

Verified a36db9cc

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

xuedinge233 committed 221 days ago

Verified c9899dc1

Lock cache file of HF model list (#6628)

tohtana committed 221 days ago

Verified 1a45bd8e

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

ShellyNR committed 221 days ago

Verified ce468c37

Support safetensors export (#6579)

xu-song committed 222 days ago

Verified bf60fc0c

Add first Step in LR Schedulers (#6597)

jomayeri committed 222 days ago

Verified 85b7469e

Accept btl_tcp_if_include option through launcher_args (#6613)

diskkid committed 222 days ago

Verified 13c16c95

Add API for updating ZeRO gradients (#6590)

tjruwase committed 222 days ago

Verified 65ab6448

[compile] Show breakdown of graph break (#6601)

delock committed 222 days ago

Verified cf41e8c4

Ignore reuse_dist_env (#6623)

tohtana committed 222 days ago

Verified 7a5bc4fd

apply fp16 autocast only to floating point values

tohtana committed 225 days ago

5c4b97f1

Add API to get devices of offload states (#6586)

tohtana committed 227 days ago

Verified adec9912

reduce setting global variables to reduce torch compile graph breaks (#6541)

NirSonnenschein committed 227 days ago

Verified d7ca3d83

AIO CPU Locked Tensor (#6592)

jomayeri committed 227 days ago

Verified a1f98bdc

Clean up prefetched parameters (#6557)

tohtana committed 227 days ago

Verified 7d751ee8

Update version.txt after 0.15.2 release (#6615)

loadams committed 227 days ago

Verified 55f7f378

Enabled Qwen2-MoE Tensor Parallelism (TP) inference (#6551)

gyou2021 committed 227 days ago

Verified 474a3288

Unpin accelerate tests, update lightning with node16 removal. (#6611)

loadams committed 227 days ago

Verified 1062a0c6

Rearrange inference OPS and stop using builder.load (#5490)

oelayan7 committed 228 days ago

Verified 645639bc

Handle when `backend` is also in compile_kwargs (#6502)

oraluben committed 228 days ago

Verified ca8b1fe9

Fix device selection using CUDA_VISIBLE_DEVICES (#6530)

tohtana committed 228 days ago

Verified 5cbbff40

Improve DS logging control (#6602)

tjruwase committed 228 days ago

Verified f74ea69a

Older