microsoft/DeepSpeed

Pull Requests Commits

Preserve selected Transformers in torch latest CI

tohtana committed 86 days ago

Verified ab9e956c

Default torch latest CI to PyTorch 2.10

tohtana committed 86 days ago

Verified ab067308

Pin Modal torch preset packages

tohtana committed 86 days ago

Verified d578ff2c

Propagate Modal torch latest dispatch env

tohtana committed 86 days ago

Verified e1a7acb9

Run Modal torch latest on manual dispatch

tohtana committed 86 days ago

Verified a58ca27b

Allow git Transformers refs in Modal torch latest CI

tohtana committed 86 days ago

Verified 3180610b

Allow git Transformers refs in torch latest CI

tohtana committed 86 days ago

Verified d0d59f44

Add Modal torch latest version presets

tohtana committed 86 days ago

Verified e52ae431

Add Transformers version input to AWS torch latest

tohtana committed 86 days ago

Verified 2f8e40b0

Install released Transformers in CPU latest workflow

tohtana committed 86 days ago

Verified 704e13ff

Add CPU torch and Transformers workflow inputs

tohtana committed 86 days ago

Verified 547cdd93

Remove stale step() docstring from DeepSpeedCPUAdam (#8011)

lucaspirola committed 86 days ago

Verified ae075141

Fix ZeRO-3 forward crash on modules with plain dict _parameters (#8009)

roycho96 committed 87 days ago

Verified d7a3972f

fix: use subprocess instead of os.system in data_analyzer.py (#7994)

orbisai0security committed 87 days ago

Verified 8cdf8651

Auto-detect CUTLASS for EvoformerAttention (#8000)

MaxTretikov committed 88 days ago

Verified d5356e07

fix(io): close aio_fd in FastFileWriter._fini to prevent fd leak (#8005)

jg-heo committed 89 days ago

Verified b01a0915

zero3: SDMA allgather via mori (sdma_allgather) (#7999)

inkcherry committed 92 days ago

Verified 66af8f03

Optimize singleton MoE collectives (#7997)

Tianyi-Franklin-Wang committed 93 days ago

Verified 55491413

Update topkgating probability-mask test expectation (#8007)

tohtana committed 93 days ago

Verified 5be4c70c

Add office hours times/link on the README (#8004)

loadams committed 94 days ago

Verified 883fefb2

Update version post-v0.19.0 release (#7996)

loadams committed 99 days ago

Verified a7811998

Update version.txt before 0.19.0 release (#7995)

loadams committed 99 days ago

Verified 98a2fbdf

[AutoSP] (Sequence Parallelism) support for Multimodal Models (ViT + LLM) (#7984)

nathon-lee committed 105 days ago

Verified 4e668fce

Add Gram Newton-Schulz orthogonalization for Muon optimizer (#7953)

delock committed 106 days ago

Verified 8a77f381

Add DeepSpeed NVTX domain support (#7988)

heurry committed 106 days ago

Verified 5b441027

fix: topkgating major bug (#7986)

excepshenal committed 106 days ago

Verified 853c9389

Fix BF16_Optimizer last-microbatch grad leak under ZeRO-1 (#7985)

maxyu1115 committed 107 days ago

Verified 5999fb06

Fix modal ci timeout (#7989)

sfc-gh-truwase committed 107 days ago

Verified 6e291bee

Dynamic offload compatible with static optimizer offload (#7979)

sfc-gh-truwase committed 112 days ago

Verified f6d7ae48

Fix Adam subgroup inconsistency (#7982)

st-bang97 committed 113 days ago

Verified 44c51e34

Older