Skip to content

Pull requests: NVIDIA/Megatron-LM

New pull request New
248 Open 1,181 Closed
248 Open 1,181 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

Remove cross-rank synchronization during checkpoint load & deprecate torch.distributed.checkpoint.state_dict_loader.load_state_dict
#2864 opened Jan 8, 2026 by asolergi-nv Loading…
1
build: Pin nvshmem
#2863 opened Jan 8, 2026 by ko3n1g Loading…
6 tasks
Change oncall team name
#2861 opened Jan 8, 2026 by Phlip79 Loading…
6 tasks
Core 0.16
2
Update Slack user group when oncall changes
#2859 opened Jan 8, 2026 by Phlip79 Loading…
6 tasks
Core 0.16
2
Cherry-pick bug fixes into 0.15.X. cherry-pick core_r0.15.0
#2858 opened Jan 7, 2026 by cspades Loading…
6 tasks
@cspades
Use global user buffer when the bucket size does not fit FixedPoolAllocator
#2857 opened Jan 7, 2026 by shengf-nv Loading…
6 tasks
1
NCCL_DEBUG
#2849 opened Jan 7, 2026 by ko3n1g Loading…
6 tasks
Core 0.16
Initialize M-FSDP Optimizer(DistOpt v2) with DTensor Parameter Support
#2848 opened Jan 7, 2026 by shjwudp • Draft
6 tasks
1
ci: Onboard GB200
#2847 opened Jan 7, 2026 by ko3n1g Loading…
6 tasks
Core 0.16
[Dev] Add Qwen3-VL support with Megatron-FSDP dev branch Dev branch related issues and development
#2842 opened Jan 7, 2026 by xuwchen Loading…
6 tasks
1
Add Qwen3-VL support with Megatron-FSDP
#2841 opened Jan 7, 2026 by xuwchen Loading…
6 tasks
1
[Dev] Support EP with HSDP community-request
#2840 opened Jan 7, 2026 by wplf Loading…
6 tasks
1
Improve memory logging
#2839 opened Jan 7, 2026 by deepakn94 Loading… Core 0.16
Support DDP overlap for models with repeated parameters
#2837 opened Jan 7, 2026 by deepakn94 Loading… Core 0.16
2
Add retry loop with exponential backoff in dataloader as a form of in-application fault tolerance complexity: low
#2836 opened Jan 7, 2026 by deepakn94 Loading… Core 0.16
5
Cherry-pick optimizer override refactor from #2723
#2835 opened Jan 6, 2026 by yaoyu-33 Loading… Core 0.16
4
Refactor spec modification/introspection to make references to Submodules typed community-request
#2834 opened Jan 6, 2026 by nschank Loading…
6 tasks
5
fsdp: avoid double sharding of MoE experts when EP is enabled community-request
#2833 opened Jan 6, 2026 by CodersAcademy006 Loading…
1
1
Supporting inference when called within an asyncio loop (Pythonic)
#2831 opened Jan 6, 2026 by tdene • Draft
6 tasks
3
Update moe readme.
#2830 opened Jan 6, 2026 by Victarry • Draft
6 tasks
1
Fix: Skip JIT warmup when fusion is disabled via arguments community-request
#2827 opened Jan 6, 2026 by kisseternity Loading…
1
[dev] feat(moe): Support apply wd to qk layernorm for Qwen3-Next
#2825 opened Jan 6, 2026 by yuzhongw-nvidia Loading…
6 tasks
Core 0.16
10
Add type hints to z_loss_func parameters community-request
#2821 opened Jan 6, 2026 by JavaZeroo Loading…
2 of 6 tasks
1
Support EP with HSDP community-request
#2820 opened Jan 6, 2026 by wplf • Draft
6 tasks
1
ci: Bump to pytorch 25.11
#2819 opened Jan 6, 2026 by chtruong814 Loading…
6 tasks
Core 0.16
ProTip! Type g i on any issue or pull request to go back to the issue listing page.