Skip to content

Pull requests: NVIDIA/Megatron-LM

New pull request New
359 Open 2,347 Closed
359 Open 2,347 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: add sync-skills workflow, rename CLAUDE.md → AGENTS.md, move .claude/skills → skills/ complexity: low
#4303 opened Apr 14, 2026 by ko3n1g Contributor Loading… Core 0.16
3
Fix aux loss computation with per-token loss and dynamic-cp
#4302 opened Apr 14, 2026 by xiaoyao0115 Contributor • Draft
5 tasks
1
Inference | Per-block MoE routing storage for prefix caching complexity: medium Final Review PR is in the "final review" stage
#4301 opened Apr 14, 2026 by lmcafee-nvidia Contributor Loading…
2 of 3 tasks
@lmcafee-nvidia
2
ci: add configurable launcher support for functional tests (ft_launcher / torchrun) complexity: low
#4298 opened Apr 14, 2026 by ko3n1g Contributor Loading…
3 tasks
Core 0.16
3
[main] thd support and GDN packed-seq alignment community-request
#4296 opened Apr 14, 2026 by DAISY-gh Contributor Loading…
3 tasks
2
Allow inference to asynchronously await GPU work completion complexity: medium
#4295 opened Apr 14, 2026 by tdene Contributor • Draft
5 tasks
Core 0.16
1
[fix] fix optimizer community-request Final Review PR is in the "final review" stage
#4294 opened Apr 14, 2026 by pavelgein Loading…
1 of 5 tasks
2
[Dev] Refactor CUDA graph configuration: separate inference granularity from training scope
#4293 opened Apr 14, 2026 by buptzyb Contributor • Draft
5 tasks
11
Refactor CUDA graph configuration: separate inference granularity from training scope
#4292 opened Apr 14, 2026 by buptzyb Contributor • Draft
5 tasks
1
chore: nightly sync main into dev (14_04_2026) complexity: high
#4291 opened Apr 14, 2026 by svcnvidia-nemo-ci Loading… Core 0.16
17
cp: NVFP4 native weights for DDP (4005) into core_r0.17.0 cherry-pick Run CICD
#4290 opened Apr 14, 2026 by ko3n1g Contributor Loading… Core 0.16
2
[Copy to main] TE fused grouped mlp with grouped bias and delayed wgrad
#4288 opened Apr 14, 2026 by svcnvidia-nemo-ci • Draft
5 tasks
2
Cuda graph fix Final Review PR is in the "final review" stage
#4285 opened Apr 13, 2026 by i-riyad Contributor Loading…
5 tasks
2
Fix assertion logic in combined_1f1b_schedule_for_interleaved_pipelining community-request
#4276 opened Apr 13, 2026 by joapolarbear Loading…
1
ci: add workflow_dispatch support to cicd-main.yml
#4275 opened Apr 13, 2026 by ko3n1g Contributor • Draft
3 tasks
1
fix mfsdp unwrap stuck at MegatronFSDP complexity: low Final Review PR is in the "final review" stage module: megatron-fsdp
#4274 opened Apr 13, 2026 by wplf Member Loading… Core 0.16
8
fix mfsdp unwrap stuck at MegatronFSDP [dev] module: megatron-fsdp
#4273 opened Apr 13, 2026 by wplf Member Loading… Core 0.16
6
[Dev] Support delayed wgrad compute overlap with P2P backward
#4268 opened Apr 13, 2026 by Wohox Contributor • Draft
5 tasks
1
[Main] Fix TE version check for retain_pinned_cpu_buffers in cpu offload complexity: low Final Review PR is in the "final review" stage
#4267 opened Apr 13, 2026 by BestJuly Contributor Loading… Core 0.16
2
Get device correctly when module returns a dict instead of individual tensor Approved All necessary approvals have been made complexity: low
#4265 opened Apr 13, 2026 by shifangx Contributor Loading…
5 tasks
4
Factor RL-specific code out of training.py complexity: high
#4264 opened Apr 12, 2026 by tdene Contributor Loading…
5 tasks
Core 0.16
4
Call save_checkpoint_and_time() when saving checkpoint and compute elapsed duration for saving checkpoint before logging timer community-request
#4263 opened Apr 12, 2026 by awsankur Loading…
3 tasks
Core 0.16
4
Add TEFusedDenseMLP for Dense+Grouped GEMM fusion on SM100+
#4262 opened Apr 12, 2026 by sraman-rgb Loading…
4 of 5 tasks
Core 0.16
14
Enable CUDA graphs for MTP inference Run functional tests
#4260 opened Apr 11, 2026 by santhnm2 Contributor • Draft
5 tasks
Core 0.16
5
New allgathervdispatcher for inference and simplify old dispatcher. complexity: high
#4258 opened Apr 10, 2026 by sidsingh-nvidia Contributor • Draft
5 tasks
Core 0.16
2
ProTip! What’s not been updated in a month: updated:<2026-03-14.