Skip to content

Pull requests: NVIDIA/TransformerEngine

New pull request New
49 Open 966 Closed
49 Open 966 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
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Make it an option to compile activation functions with fast math
#1410 opened Jan 15, 2025 by guyueh1 Loading…
1 of 13 tasks
4
Support store_param_remainders feature from Apex in TE Fused Adam
#1408 opened Jan 13, 2025 by sanandaraj5597 Loading…
[JAX] Support segment_ids/pos as FA inputs
#1406 opened Jan 13, 2025 by zlsh80826 Loading…
9 of 13 tasks
5
[JAX] Consolidate the distributed fused attention test code
#1405 opened Jan 12, 2025 by mgoldfarb-nvidia Loading…
8 of 13 tasks
9
[PyTorch] Avoid parameters function in op backward pass bug Something isn't working
#1403 opened Jan 11, 2025 by timmoon10 Loading…
3 of 13 tasks
1
Use log1p(x) instead of log(1+x)
#1401 opened Jan 11, 2025 by kit1980 Loading…
1 of 6 tasks
5
[PyTorch] Fix AttentionParams comparison logic
#1397 opened Jan 9, 2025 by cyanguwa Loading…
8 of 13 tasks
1
1
Better cuBLAS handle management
#1389 opened Jan 2, 2025 by ptrendx Loading…
8 of 13 tasks
1
@ksivaman
8
Update README.rst
#1385 opened Dec 23, 2024 by sbhavani Loading…
1 of 6 tasks
[PyTorch] Add caching for attention backend selection results
#1381 opened Dec 19, 2024 by cyanguwa • Draft
8 of 13 tasks
1
Don't touch nor send messages to the root logger.
#1380 opened Dec 19, 2024 by sagostinho-nvidia Loading…
4 of 13 tasks
1
[MoE][PyTorch] Add mask-based MoE permutation
#1373 opened Dec 13, 2024 by hxbai Loading…
8 of 13 tasks
8
[common/PyTorch] Add FusedAttention support for SWA (left, right)
#1369 opened Dec 12, 2024 by cyanguwa • Draft
8 of 13 tasks
1
Add paged attention support
#1355 opened Dec 4, 2024 by cyanguwa Loading…
8 of 13 tasks
6
[Draft] Introduce NVSHMEM based communication API for pytorch
#1346 opened Nov 22, 2024 by gdengk • Draft
13 tasks
[PyTorch] Bugfix for wgrad bulk overlap conflict when dgrad overlap is reduce-scatter bug Something isn't working
#1341 opened Nov 18, 2024 by denera Loading…
6 of 13 tasks
@denera
10
[C/JAX] Comm+GEMM Overlap API for TE/JAX enhancement New feature or request jax
#1337 opened Nov 15, 2024 by denera • Draft
3 of 13 tasks
@denera
4
Build with uv instead of just pip
#1324 opened Nov 8, 2024 by jennifgcrl Loading…
5 of 13 tasks
1
4
TP communication overlap: enable the overlap between GEMM chunk at Ho…
#1311 opened Nov 4, 2024 by erhoo82 Loading…
1 of 13 tasks
2
[JAX] Collective GEMM custom op with nvte_cublas_gemm (no comm. overlap) jax
#1307 opened Nov 2, 2024 by denera Loading…
7 of 17 tasks
@denera
14
[PyTorch] Add heuristics for intializing FP8 params enhancement New feature or request
#1300 opened Oct 30, 2024 by timmoon10 Loading…
8 of 13 tasks
2
Offloading example
#1299 opened Oct 29, 2024 by sanandaraj5597 Loading…
[PyTorch] Fix autocast deprecation warnings
#1277 opened Oct 21, 2024 by yaox12 Loading…
13 tasks
16
[PyTorch] Remove sequence parallel check for setting dropout RNG context
#1272 opened Oct 18, 2024 by ksivaman • Draft
1 of 13 tasks
attention_mask fill with -inf for UnfusedDotProductAttention
#1268 opened Oct 18, 2024 by Agoniii Loading…
1 of 13 tasks
2
ProTip! Follow long discussions with comments:>50.