-
Notifications
You must be signed in to change notification settings - Fork 628
Pull requests: NVIDIA/TransformerEngine
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
[Common] Bucket batch size with higher granularity for THD
2.13.0
#2653
opened Feb 6, 2026 by
cyanguwa
Loading…
8 of 13 tasks
[PyTorch Debug] Skip logging stats if unsupported
#2652
opened Feb 5, 2026 by
pggPL
Loading…
8 of 13 tasks
[Pytorch] Make test script generate checkpoints if they don't exist
#2650
opened Feb 4, 2026 by
kainzhong
Loading…
5 of 13 tasks
[JAX] Fix FSDP when FSDP+EP is active
#2649
opened Feb 3, 2026 by
jberchtold-nvidia
Loading…
8 of 13 tasks
Add examples for MoE models - Mixtral in TE
#2642
opened Feb 2, 2026 by
faradawn
Loading…
2 of 13 tasks
docs(readme): update convergence table, latest news, and outdated links
#2638
opened Feb 1, 2026 by
sbhavani
Loading…
5 of 13 tasks
Fix FP8 block scaling with sequence parallel
#2637
opened Jan 31, 2026 by
cuichenx
Loading…
1 of 13 tasks
[Common][PyTorch] Add a new score func
sqrtsoftplus to the fused router
MoE
#2633
opened Jan 29, 2026 by
yaox12
Loading…
4 of 16 tasks
[Common][PyTorch] Fuse scaling and unscaling of bf16 momentums into kernels
2.13.0
#2632
opened Jan 29, 2026 by
yaox12
Loading…
13 tasks
[Common] Fuse pre-swizzling into grouped MXFP8 quantization kernel
#2630
opened Jan 28, 2026 by
Oleg-Goncharov
Loading…
7 of 13 tasks
[PyTorch] Pad V when Q/V head dims differ (MLA) for THD
#2629
opened Jan 27, 2026 by
HollowMan6
Loading…
8 of 13 tasks
docs: update cuDNN sliding window attention support
#2624
opened Jan 26, 2026 by
sbhavani
Loading…
7 of 13 tasks
Previous Next
ProTip!
Add no:assignee to see everything that’s not assigned.