Skip to content

Pull requests: NVIDIA/Megatron-LM

New pull request New
254 Open 1,094 Closed
254 Open 1,094 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

Add flashinfer
#2768 opened Dec 27, 2025 by kanz-nv Loading…
6 tasks
Core 0.16
4
Do a pass of typing fixes on transformer/ community-request
#2766 opened Dec 26, 2025 by nschank Loading…
2 of 6 tasks
Core 0.16
9
Fix: Perform sigmoid calculation in fp32 for aux loss stability community-request
#2765 opened Dec 26, 2025 by CodersAcademy006 Loading…
1
1
[MoE]Enable Casting-Free FP8-Flow-MoE Blockwise FP8 Dataflow community-request
#2764 opened Dec 26, 2025 by xiaoxi-wangfj • Draft
1 of 6 tasks
1
Fuse permute+pad and unpermute+unpad ops for FP8 optimization community-request
#2763 opened Dec 26, 2025 by xiaoxi-wangfj Loading…
1 of 6 tasks
1
Replaces ModuleSpec with Protocols for some of the inputs to SelfAttention/CrossAttention community-request
#2761 opened Dec 25, 2025 by nschank Loading…
2 of 6 tasks
1
[Dev] Optimizer State and Master Weight Offloading dev branch Dev branch related issues and development
#2760 opened Dec 25, 2025 by hxbai • Draft
6 tasks
Core 0.16
@hxbai
1
Add support for UCC backend - alltoall moe layer
#2759 opened Dec 25, 2025 by ybenvidia • Draft
6 tasks
1
feat: Cherry-pick PR of PR!2661 for dev branch
#2757 opened Dec 25, 2025 by youngeunkwon0405 Loading…
6 tasks
Core 0.16
@youngeunkwon0405
2
Cleanup oncall
#2756 opened Dec 24, 2025 by Phlip79 Loading…
6 tasks
1
Only assign oncall to main PRs
#2755 opened Dec 24, 2025 by Phlip79 Loading…
6 tasks
1
[Dev] Remove calculation of padding token in moe routing loss Run functional tests
#2754 opened Dec 24, 2025 by HaochenYuan Loading…
6 tasks
Core 0.16
1
[main] feat(moe): Support apply wd to qk layernorm for Qwen3-Next (4/4)
#2753 opened Dec 24, 2025 by yuzhongw-nvidia Loading…
6 tasks
Core 0.16
4
[main] feat(moe): Support attention output gate for Qwen3-Next (3/4)
#2752 opened Dec 24, 2025 by yuzhongw-nvidia Loading…
6 tasks
Core 0.16
4
[main] feat(moe): Support moe shared expert gate for Qwen3-Next (2/4)
#2751 opened Dec 24, 2025 by yuzhongw-nvidia Loading…
6 tasks
Core 0.16
2
[Dev] Disable ep overlap memory optimization bug Something isn't working dev branch Dev branch related issues and development Expert Review Apply this label to indicate that your PR is ready for expert review.
#2750 opened Dec 24, 2025 by Wohox Loading…
6 tasks
Core 0.16
3
[DEV][WIP][DO NOT MERGE][REFACTOR] Introduce ContextParallelHandler for Unified Context Parallelism Abstraction community-request
#2749 opened Dec 24, 2025 by littsk Loading…
5 of 6 tasks
5
Add retry logic with configurable delays to checkpoint write operations complexity: medium
#2746 opened Dec 24, 2025 by apaithankar Loading…
6 tasks
3
Add RL support for MOEs Expert Review Apply this label to indicate that your PR is ready for expert review. Run functional tests Run tests
#2742 opened Dec 23, 2025 by jon-barker Loading…
6 tasks
Core 0.16
@jon-barker
7
Explicitly delete objects in async checkpoint worker loop and call GC
#2740 opened Dec 23, 2025 by apaithankar Loading…
6 tasks
Core 0.16
3
Fix Linear-Cross-Entropy Convergence Issue Expert Review Apply this label to indicate that your PR is ready for expert review.
#2739 opened Dec 23, 2025 by shjwudp Loading…
6 tasks
Core 0.16
@shjwudp
2
Dynamic Inference | Evict and re-compute context requests. Expert Review Apply this label to indicate that your PR is ready for expert review.
#2738 opened Dec 23, 2025 by lmcafee-nvidia Loading…
6 tasks
Core 0.15
@lmcafee-nvidia
6
Revert "[Dev] fix(moe): Support HybridEP and reduce memory overhead f…
#2735 opened Dec 22, 2025 by ko3n1g Loading…
6 tasks
Core 0.16
Update MoE functional test 2.0 configs
#2731 opened Dec 22, 2025 by Victarry • Draft
6 tasks
1
[Dev](moe):Refine A2A overlap under CUDA_DEVICE_MAX_CONNECTIONS=1 community-request
#2730 opened Dec 21, 2025 by Baidu-AIAK Loading…
1
ProTip! Exclude everything labeled bug with -label:bug.