Skip to content

Pull requests: NVIDIA/Megatron-LM

New pull request New
316 Open 1,880 Closed
316 Open 1,880 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

Normalize tool_calls and gate parser tool-calls to tool-enabled requests complexity: low Final Review PR is in the "final review" stage
#3710 opened Mar 4, 2026 by i-riyad Loading…
6 tasks
1
Claude to add complexity label complexity: low Expert Review Apply this label to indicate that your PR is ready for expert review.
#3709 opened Mar 4, 2026 by Phlip79 Loading…
6 tasks
Core 0.16
2
Extract the changes from Jorge's branch
#3701 opened Mar 4, 2026 by tdene • Draft
6 tasks
1
ci: Mount and enforce HF_HOME Run functional tests
#3700 opened Mar 4, 2026 by ko3n1g Loading…
6 tasks
Core 0.16
7
find optimal number of workers
#3699 opened Mar 4, 2026 by dimapihtar • Draft
6 tasks
Core 0.16
3
Fix config.softmax_scale not being considered Final Review PR is in the "final review" stage
#3698 opened Mar 4, 2026 by janEbert Loading… Core 0.16
[Main][feat] Support CUDA Graph capture offloading modules complexity: medium enhancement New feature or request
#3697 opened Mar 4, 2026 by lhb8125 • Draft
6 tasks
Core 0.16
@lhb8125
5
chore: CLI launch internal CI Expert Review Apply this label to indicate that your PR is ready for expert review.
#3695 opened Mar 4, 2026 by ko3n1g Loading…
6 tasks
Core 0.16
11
fix ddp bug when --overlap-grad-reduce and --num-distributed-optimi for dev
#3694 opened Mar 4, 2026 by wplf Loading…
6 tasks
1
fix ddp bug when --overlap-grad-reduce and --num-optim > 1 Final Review PR is in the "final review" stage
#3693 opened Mar 4, 2026 by wplf Loading…
3
Modify mfsdp default data-parallel-sharding-strategy for dev
#3692 opened Mar 4, 2026 by wplf Loading…
6 tasks
1
Modify mfsdp default data-parallel-sharding-strategy
#3691 opened Mar 4, 2026 by wplf • Draft
6 tasks
4
Add Engram model structure integration (v1)
#3689 opened Mar 4, 2026 by ilml • Draft
3 tasks
1
fix: skip FSDP DTensor boundary validation under fake process group Final Review PR is in the "final review" stage
#3686 opened Mar 4, 2026 by Victarry Loading…
6 tasks
Core 0.16
4
Fix: Defensively close GPU device FDs in dataloader worker processes
#3684 opened Mar 4, 2026 by hexinw-nvidia • Draft
4
Add doc for layerwise distributed optimizer
#3682 opened Mar 3, 2026 by BoxiangW • Draft
6 tasks
3
adding public_docs_features: True to get proper legal footer… complexity: low
#3681 opened Mar 3, 2026 by megnvidia • Queued
6 tasks
Core 0.16
6
Improve error logging when invalid number of tokens is requested. complexity: low Expert Review Apply this label to indicate that your PR is ready for expert review.
#3680 opened Mar 3, 2026 by yobibyte Loading… Core 0.16
Cuda graphed torch grouped mm for MoE inference.
#3677 opened Mar 3, 2026 by sidsingh-nvidia • Draft
6 tasks
Core 0.16
2
Fix split state dict main Final Review PR is in the "final review" stage
#3676 opened Mar 3, 2026 by kunlunl Loading…
6 tasks
1
refactor: migrate TransformerConfig validations to __post_init__ (Part of #3568) community-request Final Review PR is in the "final review" stage
#3675 opened Mar 3, 2026 by CodersAcademy006 Loading…
1
4
Enable DSA CP/absorbed/THD paths with TileLang fused ops community-request
#3674 opened Mar 3, 2026 by HollowMan6 • Draft
6 tasks done
14
[MoE] Add a new score function to the router Run functional tests
#3673 opened Mar 3, 2026 by yaox12 • Draft
6 tasks
Core 0.16
2
Fix split_state_dict function for MoE models community-request Expert Review Apply this label to indicate that your PR is ready for expert review.
#3667 opened Mar 3, 2026 by eternally-z Loading…
6 tasks
Core 0.16
7
Minor inference changes for NemoRL
#3666 opened Mar 3, 2026 by ArEsKay3 • Draft
6 tasks
1
ProTip! Exclude everything labeled bug with -label:bug.