Skip to content

Pull requests: NVIDIA/Megatron-LM

New pull request New
766 Open 3,832 Closed
766 Open 3,832 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

ci(auth): treat svcnemo-autobot as internal complexity: low
#6437 opened Aug 11, 2026 by ko3n1g Contributor Loading…
1
ci(auth): treat svcnemo-autobot as internal complexity: low
#6436 opened Aug 11, 2026 by ko3n1g Contributor Loading…
2
ci: AUT-1408 backport GitHub App authentication to dev
#6435 opened Aug 11, 2026 by svcnemo-autobot Collaborator Loading…
2
docs: add configuration errors troubleshooting page community-request
#6434 opened Aug 11, 2026 by erensh27 • Draft
1
2
cp: build: bump transformer-engine to 2.17.1 Run tests
#6433 opened Aug 11, 2026 by svcnemo-autobot Collaborator • Draft
3
Avoid unnecessary MoE router host synchronization complexity: low nemotron
#6432 opened Aug 11, 2026 by JF-D Contributor Loading…
6 tasks done
1
@JF-D
4
Add MFSDP v2 design document docs-only documentation only (docs or docstrings) Final Review PR is in the "final review" stage
#6431 opened Aug 11, 2026 by wujingyue Contributor Loading…
1
Add async scheduling pairwise unit coverage
#6430 opened Aug 11, 2026 by lmcafee-nvidia Contributor • Draft
1
[Main] Recompute EP A2A overlap activations per layer segment (supersedes #5920)
#6429 opened Aug 11, 2026 by Wohox Contributor • Draft
5 of 6 tasks
1
Test GPTOSS SSE with async scheduling Run functional tests
#6428 opened Aug 11, 2026 by lmcafee-nvidia Contributor • Draft
1
Kimi K3 VPP + cross-stage caching community-request
#6427 opened Aug 11, 2026 by wuweiqiang24 • Draft
6 tasks
1
Add DSA over GQA with the streamed min-memory kernels complexity: high
#6426 opened Aug 11, 2026 by alokpathy Loading…
3 of 6 tasks
1
Add initial Muon+M-FSDPv2 implementation
#6425 opened Aug 11, 2026 by janEbert Contributor • Draft
1 task
4
Add multimodal tokenizer prompt formats complexity: high Run tests
#6424 opened Aug 10, 2026 by matthieule Contributor Loading…
4 of 6 tasks
3
Propagate RL runtime config updates to module configs complexity: low
#6423 opened Aug 10, 2026 by Phlip79 Member Loading…
1 task done
2
Add tensor-parallel Muon Hyperball optimizer
#6422 opened Aug 10, 2026 by mkhona-nvidia Contributor • Draft
3
build(container): support non-root CI execution Run tests
#6421 opened Aug 10, 2026 by ko3n1g Contributor • Draft
1
Paged stashing no longer assumes single layer config complexity: low
#6419 opened Aug 10, 2026 by Phlip79 Member Loading…
1 task done
2
Add prefix cache stress test foundation
#6418 opened Aug 10, 2026 by lmcafee-nvidia Contributor • Draft
1
Preserve greedy temperature in RL inference
#6417 opened Aug 10, 2026 by lmcafee-nvidia Contributor • Draft
2 of 6 tasks
1
Defer prefix-cache hit accounting until admission succeeds
#6416 opened Aug 10, 2026 by lmcafee-nvidia Contributor • Draft
2 of 6 tasks
1
Add opt-in CuTeDSL copy kernels for MoE paged stash
#6414 opened Aug 10, 2026 by sraman-rgb Contributor • Draft
6 tasks
2
Fix EP ZMQ init ordering before CUDA graph capture community-request
#6412 opened Aug 10, 2026 by YigongQin • Draft
5 of 6 tasks
1
Fix tool call reasoning boundary complexity: low
#6411 opened Aug 10, 2026 by nvcsathe Contributor Loading…
6 tasks
1
Add hybrid layer config definitions complexity: low Final Review PR is in the "final review" stage
#6410 opened Aug 10, 2026 by Phlip79 Member Loading…
1 task done
2
ProTip! Type g i on any issue or pull request to go back to the issue listing page.