Skip to content

Pull requests: deepspeedai/DeepSpeed

New pull request New
144 Open 3,735 Closed
144 Open 3,735 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Release GIL in blocking AIO bindings
#8196 opened Jul 30, 2026 by tohtana Collaborator Loading…
4
MoE dispatch: fix silent Tutel + TP corruption, speed up native and Tutel paths
#8195 opened Jul 29, 2026 by jinyouzhi Contributor Loading…
3
Derive AutoEP rank splits from the per-expert count exchange
#8190 opened Jul 29, 2026 by yh0903 Loading…
@tohtana
7
Enable ZeRO-3 linear wrapper for existing models
#8189 opened Jul 28, 2026 by tohtana Collaborator Loading…
10
Remove redundant DCO workflow
#8188 opened Jul 28, 2026 by tohtana Collaborator • Draft
4
Wiring up offload_opt_states
#8186 opened Jul 28, 2026 by pengdurice Contributor • Draft
[WIP][AutoTP] Complete uneven sharding and universal checkpoint support
#8185 opened Jul 28, 2026 by jinyouzhi Contributor Loading…
15
Add managed_gradient_accumulation for ZeRO stage 0/1
#8184 opened Jul 27, 2026 by sfc-gh-truwase Collaborator Loading…
1
19
Add keyword argument support to activation checkpointing
#8182 opened Jul 27, 2026 by zupengwang • Draft
1
Non-reentrant activation checkpoint CPU offload
#8181 opened Jul 27, 2026 by winglian Loading…
6
Add Triton grouped-GEMM for MoE experts on Ampere/Ada
#8180 opened Jul 26, 2026 by hwchen2017 Contributor Loading…
18
Make PipelineModule.set_checkpoint_interval actually change the interval
#8178 opened Jul 25, 2026 by vineethsaivs Contributor Loading…
4
Stabilize DeepCompile ZeRO-3 memory scheduling
#8169 opened Jul 23, 2026 by tohtana Collaborator • Draft
[AutoTP] Fix ZeRO-3 checkpoint consolidation to gather across TP and DP
#8168 opened Jul 23, 2026 by delock Collaborator Loading…
12
fix: avoid ZeroDivisionError in compute_elastic_config return_microbatch path
#8162 opened Jul 21, 2026 by Solaris-star Loading…
1
[Pipeline] Add Ray-backed stage placement and execution with pluggable transportsFeature/ray pipeline parallelism
#8161 opened Jul 21, 2026 by SisPiao Loading…
6
Fix ZeRO-3 all_reduce param fetch stride for padded parameters
#8158 opened Jul 19, 2026 by ebarkhordar Contributor Loading…
8
Fix DeepCompile ZeRO-3 gathered parameter ownership
#8157 opened Jul 19, 2026 by tohtana Collaborator Loading…
10
Clamp capacity to num_tokens in MoE gating drop branches (complete #5353)
#8155 opened Jul 18, 2026 by ebarkhordar Contributor Loading…
7
Add governance process page to deepspeed.ai
#8138 opened Jul 13, 2026 by PKUWZP Collaborator Loading…
3 tasks
5
Feature/zero multi loss separate backward
#8135 opened Jul 12, 2026 by nathon-lee Contributor Loading…
14
fix: add zero-guards for divisor/scale edge cases (#7838)
#8133 opened Jul 12, 2026 by liuyun7345 Loading…
7 tasks done
1
2
docs: add CHANGELOG.md
#8116 opened Jul 3, 2026 by NovusEdge Contributor Loading…
1
3
Add Ulysses SP support for FLA gated-delta context parallelism
#8114 opened Jul 2, 2026 by xylian86 Collaborator Loading…
4
Fix formatting CI failures in benchmarks/opsd benchmark scripts
#8111 opened Jul 1, 2026 by Copilot AI • Draft
@PKUWZP
1
ProTip! Follow long discussions with comments:>50.