Skip to content

Pull requests: NVIDIA/TransformerEngine

New pull request New
122 Open 2,065 Closed
122 Open 2,065 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 head dim 256 support for SDPA on Blackwell
#2906 opened Apr 21, 2026 by yaox12 Member Loading…
1 of 13 tasks
6
[PyTorch] Fix cuteDSL kernel incorrect numerics when K is 64 aligned 2.15.0
#2905 opened Apr 21, 2026 by ksivaman Member Loading…
6 of 13 tasks
4
Make NS coefficients parameter 2D in Python API 2.15.0
#2904 opened Apr 20, 2026 by vcherepanov-nv Collaborator Loading…
5 of 13 tasks
2
Fix CP crash with GQA + asymmetric KV head dims (#2868) community-contribution PRs from external contributor outside the core maintainers, representing community-driven work.
#2901 opened Apr 19, 2026 by beccohov Loading…
7 of 13 tasks
1
@cyanguwa
2
[PyTorch] Expose function to bulk-allocate tensors backed by the same buffer
#2900 opened Apr 18, 2026 by timmoon10 Collaborator Loading…
9 of 13 tasks
5
add support for enabling cuda graph under thd format in megatron.
#2898 opened Apr 17, 2026 by HaochenYuan Loading…
13 tasks
5
Improve the dimension checks for the FP8 recipes
#2894 opened Apr 16, 2026 by ptrendx Member Loading…
13 tasks
16
Add optimised top-k kernel AIR.
#2890 opened Apr 16, 2026 by dcampora Loading…
8 of 13 tasks
@ptrendx
17
Add AI written qwen3_moe example
#2887 opened Apr 15, 2026 by skyw Loading…
4 of 13 tasks
@pggPL
3
Bias/Dbias Support for GroupedLinear
#2885 opened Apr 15, 2026 by vthumbe1503 Collaborator Loading…
13 tasks
27
[Debug] Add AutoswitchGEmm for Debug Precision Tool
#2883 opened Apr 15, 2026 by shangxiaokang • Draft
3 of 13 tasks
@pggPL
SMEM offset caching RHT
#2882 opened Apr 15, 2026 by sraman-rgb Loading…
13 tasks
7
fix(readme): update broken links and modernize project description
#2879 opened Apr 14, 2026 by sbhavani Collaborator Loading…
3 of 13 tasks
3
[PyTorch] Split TE ops op_forward into op_forward and setup_context
#2877 opened Apr 14, 2026 by pggPL Collaborator Loading…
5 of 7 tasks
12
[DONOT MERGE] Wgrad cute dsl v2
#2872 opened Apr 13, 2026 by vthumbe1503 Collaborator • Draft
13 tasks
2
[JAX] Add debug validation mode for runtime group size alignment
#2867 opened Apr 11, 2026 by jberchtold-nvidia Collaborator • Draft
13 tasks
4
Optimizations for MXFP8/NVFP4 dequantize kernels
#2865 opened Apr 10, 2026 by YigongQin Loading…
8 of 13 tasks
@Oleg-Goncharov
8
Adds GEMM Profiling Guide to TE
#2863 opened Apr 9, 2026 by jomitchellnv Contributor Loading…
7 tasks
5
[DO NOT MERGE] Test CI
#2862 opened Apr 9, 2026 by cyanguwa Collaborator • Draft
13 tasks
Add cpplint and ruff linter to pre-commit and fix lint violations
#2853 opened Apr 8, 2026 by pstjohn Contributor Loading…
@ksivaman
7
Bump transformers from 4.55.0 to 5.0.0rc3 in /docs/examples/te_gemma dependencies Pull requests that update a dependency file python Pull requests that update python code
#2851 opened Apr 8, 2026 by dependabot bot Loading…
@sudhakarsingh27
Bump transformers from 4.57.0 to 5.0.0rc3 in /docs/examples/te_llama dependencies Pull requests that update a dependency file python Pull requests that update python code
#2850 opened Apr 8, 2026 by dependabot bot Loading…
@sudhakarsingh27
Skip activation kernels when tensor size is zero bug Something isn't working
#2848 opened Apr 8, 2026 by timmoon10 Collaborator Loading…
8 of 13 tasks
11
[Common] Multicast Fixes
#2847 opened Apr 8, 2026 by phu0ngng Collaborator • Draft
13 tasks
3
[Core] Report CUDA versions when NVRTC compilation fails enhancement New feature or request
#2842 opened Apr 7, 2026 by timmoon10 Collaborator Loading…
8 of 13 tasks
1
15
ProTip! Follow long discussions with comments:>50.