-
Notifications
You must be signed in to change notification settings - Fork 5k
Pull requests: deepspeedai/DeepSpeed
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
[Workflow] run modal GPU workflows only from the merge queue
#8412
opened Sep 4, 2026 by
delock
Collaborator
Loading…
fix: validate max_out_tokens and min_out_tokens are positive
#8410
opened Sep 4, 2026 by
KaiOnCode
Loading…
Gate the offload-state memory deltas on allocator-backed stats
#8409
opened Sep 4, 2026 by
delock
Collaborator
Loading…
Forward the inputs DSUNet._forward accepts to the UNet
#8408
opened Sep 3, 2026 by
vineethsaivs
Contributor
Loading…
Probe the device module for train_cifar's fork_rng device entries
#8407
opened Sep 3, 2026 by
delock
Collaborator
Loading…
Checkpoint the curriculum data sampler's own RNG state
#8406
opened Sep 3, 2026 by
ebarkhordar
Contributor
Loading…
Add remaining parallelism rank and world-size engine APIs
#8402
opened Sep 3, 2026 by
YeonwooSung
Loading…
Log ZeRO and parallelism settings to monitor backends
#8401
opened Sep 3, 2026 by
YeonwooSung
Loading…
Fix ZeRO checkpoint DP merge on scalar optimizer state
#8396
opened Sep 2, 2026 by
Kayvan-Zahiri
Loading…
Size the elastic v0.2 micro batch from the data-parallel world
#8395
opened Sep 2, 2026 by
vineethsaivs
Contributor
Loading…
Size the interpolate flops from the whole output tensor
#8394
opened Sep 2, 2026 by
vineethsaivs
Contributor
Loading…
Fix-FP16-ZeRO-Stage-0-loss-Scaling-In-engine.backward()
#8393
opened Sep 2, 2026 by
etemadiamd
Loading…
Persist ZeRO-3 HybridMegatron QKV layout changes
#8392
opened Sep 2, 2026 by
LiRunGuo
Contributor
Loading…
Offload a saved view when it is the last value holding its storage
#8388
opened Sep 1, 2026 by
pengdurice
Contributor
•
Draft
Use the val argument in mask_nan_or_inf_with_val_inplace
#8386
opened Sep 1, 2026 by
vineethsaivs
Contributor
Loading…
Describe universal checkpoint shards as affine maps
#8385
opened Sep 1, 2026 by
Achyuthan-S
Contributor
Loading…
[muon] Per-head Newton-Schulz for attention projections
#8384
opened Sep 1, 2026 by
alanhuangyoo
Contributor
Loading…
Fix silent gradient-bucket drop on CPU in ZeRO-1/2 non-contiguous reduction
#8382
opened Sep 1, 2026 by
delock
Collaborator
Loading…
[DON'T MERGE] Run multi-rank CPU unit tests in CI via LOCAL_SIZE
#8381
opened Sep 1, 2026 by
delock
Collaborator
Loading…
[AutoEP] Make router weight persistent to avoid 0 tensor size crash
#8379
opened Aug 31, 2026 by
pengdurice
Contributor
Loading…
Previous Next
ProTip!
What’s not been updated in a month: updated:<2026-08-04.