Skip to content

Pull requests: NVIDIA/TransformerEngine

New pull request New
182 Open 2,500 Closed
182 Open 2,500 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

[PyTorch] torch.compile support for FusedAttention
#3472 opened Sep 3, 2026 by pggPL Collaborator • Draft
7 of 13 tasks
[PyTorch] DeepSeekV3Layer: full MoE transformer layer (MLA + DeepSeek MoE)
#3471 opened Sep 3, 2026 by pggPL Collaborator • Draft
7 of 13 tasks
[PyT] Guard THD learnable dSink on older cuDNN
#3470 opened Sep 3, 2026 by KshitijLakhani Collaborator Loading…
1 of 13 tasks
1
@KshitijLakhani
3
[PyTorch] Fix FP8 illegal memory access in single-process multi-GPU execution community-contribution PRs from external contributor outside the core maintainers, representing community-driven work.
#3469 opened Sep 3, 2026 by SuperGoodGame Loading…
1
2
Add opt-in rowwise-only quantized primary weights community-contribution PRs from external contributor outside the core maintainers, representing community-driven work.
#3468 opened Sep 3, 2026 by xiuhu17 Contributor Loading…
8 of 13 tasks
1
[Common] row-scaled nvfp4 path: add single-launch group fused amax community-contribution PRs from external contributor outside the core maintainers, representing community-driven work.
#3467 opened Sep 3, 2026 by cael-ling Contributor Loading…
2 of 13 tasks
3
[PyTorch] Fix: Resolve EP symm-mem window offset for both old and new torch version
#3466 opened Sep 2, 2026 by phu0ngng Collaborator Loading…
7 of 13 tasks
3
Compress CUDA binaries
#3465 opened Sep 2, 2026 by fheinecke Collaborator • Draft
1 of 13 tasks
1
[JAX] Make MoEBlock aware of Dense TP axes
#3462 opened Sep 2, 2026 by jberchtold-nvidia Collaborator • Draft
8 of 13 tasks
Add non-TMA MXFP8 cast-only kernels for specialized rowwise-only and row+colwise
#3459 opened Sep 1, 2026 by tdophung Collaborator • Draft
13 tasks
1
[Common] Group NVFP4 Quantize Kernels
#3458 opened Sep 1, 2026 by Oleg-Goncharov Collaborator Loading…
9 of 13 tasks
4
[JAX] Add SiTU-GLU support to JAX
#3457 opened Sep 1, 2026 by jberchtold-nvidia Collaborator • Draft
8 of 13 tasks
1
[PyTorch] Schedule delayed-scaling updates after backward
#3456 opened Sep 1, 2026 by pggPL Collaborator Loading…
9 of 13 tasks
1
[Common] row-scaled nvfp4 path: fuse row/col amax into a single TMA-tiled kernel community-contribution PRs from external contributor outside the core maintainers, representing community-driven work.
#3454 opened Sep 1, 2026 by cael-ling Contributor Loading…
2 of 13 tasks
2
[JAX] Add EP borrowed-comm path via XLA FFI collectives extension
#3452 opened Aug 31, 2026 by phu0ngng Collaborator • Draft
13 tasks
[JAX] Add sqrtsoftplus router score function
#3448 opened Aug 31, 2026 by jberchtold-nvidia Collaborator Loading…
8 of 13 tasks
19
[JAX] Attention support for explicit sharding
#3446 opened Aug 31, 2026 by jberchtold-nvidia Collaborator • Draft
8 of 13 tasks
1
Cutlass Upgrade to 4.8 2.20
#3443 opened Aug 31, 2026 by vthumbe1503 Collaborator Loading…
13 tasks
3
Relax runtime checks for activation recompute into Warnings community-contribution PRs from external contributor outside the core maintainers, representing community-driven work.
#3436 opened Aug 28, 2026 by ghadiaravi13 Contributor • Draft
1 of 13 tasks
2
[PyTorch] Reduce CUDA graph memory retention community-contribution PRs from external contributor outside the core maintainers, representing community-driven work.
#3427 opened Aug 26, 2026 by buptzyb Contributor Loading…
@ksivaman
12
[PyTorch] Fix mutable QB bounds in CUDA graphs org-contribution
#3426 opened Aug 26, 2026 by harryzhou2000 Contributor Loading…
@denera
6
Support paged stashing for GroupedLinear activations community-contribution PRs from external contributor outside the core maintainers, representing community-driven work.
#3423 opened Aug 25, 2026 by lhb8125 Contributor Loading…
1
[PyTorch] Allow CP P2P transport group overrides community-contribution PRs from external contributor outside the core maintainers, representing community-driven work.
#3420 opened Aug 24, 2026 by xiaoyao0115 • Draft
[Common] Fix TMA synchronization in quantization kernels
#3417 opened Aug 22, 2026 by Oleg-Goncharov Collaborator Loading…
6 of 13 tasks
3
Ring Attention: free unused kv comm buffers community-contribution PRs from external contributor outside the core maintainers, representing community-driven work.
#3411 opened Aug 21, 2026 by francesco-bertolotti Contributor Loading…
6 of 13 tasks
2
ProTip! Follow long discussions with comments:>50.