- 03 Nov, 2025 1 commit
-
-
zhaochao authored
-
- 30 Oct, 2025 3 commits
-
-
Kshitij Lakhani authored
[PyT] Bump the min version expected to supported FP8 current scaling determinism on Blackwell (#2316) * Bump the min version expected to supported FP8 cs det on Blackwell Signed-off-by:
Kshitij Lakhani <klakhani@nvidia.com> * Disable fused attn for cudnn < 9.14 for FP8 CS. Disable fused attn for cudnn < 9.18 for FP8 deterministic CS Signed-off-by:
pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
-
Kirthi Shankar Sivamani authored
* Fix attention backend and tests for sm120 Signed-off-by:
Kirthi Shankar Sivamani <ksivamani@nvidia.com> * Disable MLA only for backward Signed-off-by:
-
Kirthi Shankar Sivamani authored
* Fix CI failures due to deterministic attention Signed-off-by:
-
- 28 Oct, 2025 1 commit
-
-
Charlene Yang authored
* add max_score for fused/unfused F16 non-CP Signed-off-by:
Charlene Yang <8636796+cyanguwa@users.noreply.github.com> * calculate max per head instead of max over all heads Signed-off-by:
-
- 24 Oct, 2025 1 commit
-
-
Kirthi Shankar Sivamani authored
* Support wheel build for cuda 13 Signed-off-by:
-
- 17 Oct, 2025 1 commit
-
-
dongcl authored
-
- 16 Oct, 2025 3 commits
-
-
xiaoxi-wangfj authored
* [PyTorch] Add record_stream and untyped_storage func op in QuantizedTensor Signed-off-by:
xiaoxi-wangfj <690912414@qq.com> * Update transformer_engine/pytorch/tensor/float8_blockwise_tensor.py Co-authored-by:
Tim Moon <4406448+timmoon10@users.noreply.github.com> Signed-off-by:
-
Selvaraj Anandaraj authored
Signed-off-by:
Selvaraj Anandaraj <selvaraja@login-ptyche02.ptyche.clusters.nvidia.com> Co-authored-by:
-
yuguo authored
-
- 14 Oct, 2025 3 commits
-
-
Tim Moon authored
* Require cuDNN 9.14.0+ for fused attention with FP8 current scaling Signed-off-by:
Tim Moon <tmoon@nvidia.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci --------- Signed-off-by:
-
Kirthi Shankar Sivamani authored
* Initial API change Signed-off-by:
-
Evgeny Tsykunov authored
* Fix update_quantized in ref nvfp4 quantizer Signed-off-by:
Evgeny <etsykunov@nvidia.com> * Subclass quantization API Signed-off-by:
-
- 13 Oct, 2025 3 commits
-
-
Selvaraj Anandaraj authored
* FSDP grad fusion support Signed-off-by:
Selvaraj Anandaraj <selvaraja@login-ptyche01.ptyche.clusters.nvidia.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Re-factored grad overwriting usage Signed-off-by:
Selvaraj Anandaraj <selvaraja@nvidia.com> * Update transformer_engine/pytorch/ops/fused/backward_linear_add.py Co-authored-by:
-
Peter St. John authored
Signed-off-by:
Peter St. John <pstjohn@nvidia.com> Co-authored-by:
-
Selvaraj Anandaraj authored
* Added multi-layout support for attention Signed-off-by:
Pawel Gadzinski <pgadzinski@nvidia.com> --------- Signed-off-by:
Paweł Gadziński <62263673+pggPL@users.noreply.github.com>
-
- 09 Oct, 2025 4 commits
-
-
Peter St. John authored
Don't pickle an empty dict in LayerNorm and BasicOperation layers Signed-off-by:
-
Kirthi Shankar Sivamani authored
* Update minimum python version to 3.10 and update CI Signed-off-by:
-
dongcl authored
-
Kirthi Shankar Sivamani authored
Deprecate old float8_tensor.py Signed-off-by:
-
- 08 Oct, 2025 1 commit
-
-
Kirthi Shankar Sivamani authored
Unblock path for fusing NVFP4 quantize and bgrad Signed-off-by:
-
- 04 Oct, 2025 1 commit
-
-
Kirthi Shankar Sivamani authored
* Fix in FP8 attention selection logic Signed-off-by:
-
- 03 Oct, 2025 1 commit
-
-
Jan Bielak authored
* Update to_string(NVTEScalingMode) to include block scaling Signed-off-by:
Jan Bielak <jbielak@nvidia.com> * Add `nvte_swizzle_block_scaling_to_mxfp8_scaling_factors` Signed-off-by:
Copilot <175728472+Copilot@users.noreply.github.com> Signed-off-by:
-
- 02 Oct, 2025 2 commits
-
-
hx authored
* fix overflow of int32 in permute kernels Signed-off-by:
Hongxiao Bai <hongxiaob@nvidia.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci --------- Signed-off-by:
Xin Yao <xiny@nvidia.com> Co-authored-by:
-
Tim Moon authored
* Make sure to set usages for linear op quantizers before forward Signed-off-by:
-
- 01 Oct, 2025 3 commits
-
-
Przemyslaw Tredak authored
* Fix the cublas workspace alignment Signed-off-by:
Przemek Tredak <ptredak@nvidia.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Fix Co-authored-by:
Przemyslaw Tredak <ptrendx@gmail.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci --------- Signed-off-by:
-
Evgeny Tsykunov authored
Fix QuantizedTensorBase -> QuantizedTensorStorage Signed-off-by: -
Evgeny Tsykunov authored
* Introduce QuantizerBase Signed-off-by:
-
- 30 Sep, 2025 2 commits
-
-
vthumbe1503 authored
* Test working as I think it should work Signed-off-by:
Varun Thumbe <vthumbe@nvidia.com> [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci Signed-off-by:
Vladimir Cherepanov <vcherepanov@nvidia.com> * Saving... Signed-off-by:
Vladimir Cherepanov <56651474+mk-61@users.noreply.github.com> * Add license Signed-off-by:
Ming Huang <mingh@nvidia.com> * Slightly refactor Signed-off-by:
Phuong Nguyen <phuonguyen@nvidia.com> Signed-off-by:
Jeremy Berchtold <jberchtold@nvidia.com> Co-authored-by:
jberchtold-nvidia <158520091+jberchtold-nvidia@users.noreply.github.com> Signed-off-by:
zhongboz <zhongboz@nvidia.com> Co-authored-by:
oliver könig <okoenig@nvidia.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Update setup.py Signed-off-by:
Robin Zhang <robinz@nvidia.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * update Signed-off-by:
Vasudevan Rengasamy <vrengasamy@nvidia.com> Co-authored-by:
Selvaraj Anandaraj <selvaraja@login-preos01.a51.clusters.nvidia.com> * Fixed typo Signed-off-by:
djns99 <40156487+djns99@users.noreply.github.com> Signed-off-by:
Casper <casperbh.96@gmail.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci Signed-off-by:
Hongbin Liu <hongbinl@nvidia.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * fix CI Signed-off-by:
tongliu <tongliu@nvidia.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci --------- Signed-off-by:
Jonathan Mitchell <jomitchell@login-eos02.eos.clusters.nvidia.com> * assert line change Signed-off-by:
Sudhakar Singh <sudhakars@nvidia.com> * address review comments Signed-off-by:
Yuzhong Wang <yuzhongw@nvidia.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Update transformer_engine/pytorch/tensor/_internal/float8_blockwise_tensor_base.py Signed-off-by:
Min Yang <min.yang@shopee.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * refactor: refactor multi tensor gemm interface Signed-off-by:
alan yang <89962857+cassiewilliam@users.noreply.github.com> Co-authored-by:
zhujian <zhujian.whu.cs@gmail.com> Fix cuDNN version checks when getting backend and for sm89 kv cache (#2185) * Fix cudnn version checks for kv cache for sm89. Add cudnn version check in preparation for 9.14 when getting backend Signed-off-by:
-
Charlene Yang authored
* debug existing usage Signed-off-by:
-
- 29 Sep, 2025 1 commit
-
-
Kirthi Shankar Sivamani authored
* Add NVFP4 recipe Signed-off-by:
Frank Sun <frsun@nvidia.com> Co-authored-by:
Oleg Goncharov <ogoncharov@nvidia.com> Co-authored-by:
Teddy Do <tdophung@nvidia.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Add MathDx dependency to GitHub builds Signed-off-by:
-
- 22 Sep, 2025 1 commit
-
-
Charlene Yang authored
* first draft; debug plan failure Signed-off-by:
Chen Cui <chcui@nvidia.com> Signed-off-by:
-
- 19 Sep, 2025 1 commit
-
-
yuguo authored
-
- 18 Sep, 2025 6 commits
-
-
Kshitij Lakhani authored
* Fix cudnn version checks for kv cache for sm89. Add cudnn version check in preparation for 9.14 when getting backend Signed-off-by:
-
wenjh authored
Signed-off-by:wenjh <wenjh@sugon.com>
-
wenjh authored
Signed-off-by: -
zhujian authored
feature(FA3,MLA,CP): 1. Update FA3 to commit-id 3ba6f82 (tag 2.8.0.post2 with compile error fixed), PR-1604 support hdimQK != hdimV backward 2. Update get_attention_backend method because FA3 support MLA now 3. Add CP MLA support for FA3 4. Add unit tests for FA3 MLA CP 5. Update attention doc Signed-off-by: -
yuguo authored
-
alan yang authored
* feat: add cutlass group gemm support Signed-off-by:
-
- 17 Sep, 2025 1 commit
-
-
yuzhongw-nvidia authored
* fix memory overhead of all gather from sequence parallel Signed-off-by:
-
