microsoft/DeepSpeed

Pull Requests Commits

Version fix to unblock pypi (#8039)

sfc-gh-truwase committed 24 days ago

Verified d5d118c2

Fix DeepCompile all-gather scheduler candidate selection (#8033)

tohtana committed 24 days ago

Verified 819af0e5

Simplify module_inject.transpose (#8028)

xbcReal committed 25 days ago

Verified 1afa3a53

feat(zero): enable torch.func transforms on engine for ZeRO 0/1/2 (#8026)

roycho96 committed 26 days ago

Verified 4370aa50

Add engine.coalesce_grad_reduction() for ZeRO 1/2/3 multi-backward (#7992)

roycho96 committed 26 days ago

Verified 60b242af

Bump version (#8030)

sfc-gh-truwase committed 26 days ago

Verified a0b98d34

Keep required CI checks visible for ignored paths (#8019)

tohtana committed 27 days ago

Verified 77a249f1

Fix full CI test isolation for ZeRO chmod and NVMe quantization tests (#8008)

tohtana committed 27 days ago

Verified 4570c508

Fix PR-target workflow concurrency groups (#8017)

tohtana committed 27 days ago

Verified b7aef4dc

Run AutoSP compile tests sequentially (#8020)

tohtana committed 27 days ago

Verified 2c8a007b

fix(fp16): filter requires_grad in FP16 optimizer flat buffer init (#8029)

avicooper1 committed 27 days ago

Verified 01e3a28b

Fix DeepCompile AOT kwargs patching for PyTorch >= v2.11 (#8024)

tohtana committed 29 days ago

Verified 510ebe58

Support flash-attn 2.7.0 in FPDT attention (#8022)

xbcReal committed 30 days ago

Verified 45429221

fix(zero): enable vmap on LinearFunctionForZeroStage3 (#8023)

roycho96 committed 30 days ago

Verified ae576f83

Sort and dedupe -gencode flags emitted by op_builder.builder (#8021)

adityasingh2400 committed 31 days ago

Verified 43166463

fix: add setup_context for torch.func compatibility (#7916)

roycho96 committed 33 days ago

Verified 67b9e212

fix gemma4 num attention head bugs (from #7975) (#7990)

delock committed 33 days ago

Verified bd71eed6

[Blog] Muon Optimizer Support in DeepSpeed (#7962)

delock committed 33 days ago

Verified de473091

[fix] fix test_zf.py hang bug (#8012)

xbcReal committed 34 days ago

Verified 26ee3856

Support bf16 optimizer states with CPU offload (#8010)

lucaspirola committed 34 days ago

Verified 3c337b54

Make GitHub Actions job names unique (#8014)

tohtana committed 34 days ago

Verified 9f95452e

Run FastFileWriter fd-close test outside pytest-forked (#8015)

tohtana committed 34 days ago

Verified d6d4b3d0

Add configurable torch-latest dependency versions (#8016)

tohtana committed 34 days ago

Verified b387166f

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

lucaspirola committed 35 days ago

Verified ae075141

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

roycho96 committed 35 days ago

Verified d7a3972f

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

orbisai0security committed 35 days ago

Verified 8cdf8651

Auto-detect CUTLASS for EvoformerAttention (#8000)

MaxTretikov committed 36 days ago

Verified d5356e07

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

jg-heo committed 37 days ago

Verified b01a0915

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

inkcherry committed 40 days ago

Verified 66af8f03

Optimize singleton MoE collectives (#7997)

Tianyi-Franklin-Wang committed 41 days ago

Verified 55491413

Newer Older