- 04 Oct, 2025 1 commit
-
-
Kirthi Shankar Sivamani authored
* Fix in FP8 attention selection logic Signed-off-by:
Kirthi Shankar Sivamani <ksivamani@nvidia.com> * Improve 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:
Tim Moon <4406448+timmoon10@users.noreply.github.com> * Apply suggestions from code review Signed-off-by:
pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-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:
Tim Moon <tmoon@nvidia.com> * Avoid unsupported case for fused dbias+quantize kernel Hopper does not support dbias + FP8 cast without FP8 transpose. 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 <etsykunov@nvidia.com>
-
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:
Kshitij Lakhani <klakhani@nvidia.com> Signed-off-by:
Charlene Yang <8636796+cyanguwa@users.noreply.github.com> * fix remaining CI failures Signed-off-by:
Pawel Gadzinski <pgadzinski@nvidia.com> * fix 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-ptyche01.ptyche.clusters.nvidia.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Fixed typo Signed-off-by:
Selvaraj Anandaraj <selvaraja@login-preos01.a51.clusters.nvidia.com> * Fixed typo Signed-off-by:
Paweł Gadziński <62263673+pggPL@users.noreply.github.com> 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:
-
- 18 Sep, 2025 3 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:
-
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: -
alan yang authored
* feat: add cutlass group gemm support Signed-off-by:
-
- 17 Sep, 2025 3 commits
-
-
yuzhongw-nvidia authored
* fix memory overhead of all gather from sequence parallel Signed-off-by:
-
Sudhakar Singh authored
* add tutorial files and other local changes Signed-off-by:
-
vthumbe1503 authored
* Test working as I think it should work Signed-off-by:
-
- 16 Sep, 2025 1 commit
-
-
Jan Bielak authored
* Do not use norm fwd + amax fusion if cudnn backend is requested Signed-off-by:
-
- 15 Sep, 2025 1 commit
-
-
Tim Moon authored
* Add CPU offloading logic to ops. Fix test to compute dgrad. Signed-off-by:
-
- 10 Sep, 2025 1 commit
-
-
jomitchellnv authored
Adds context parallelism utilities: moving cp shards to diff ranks and pad sequence to divisibility factory (#2129) * test - adds unit test for cp utilities and the utilites Signed-off-by:
-
- 08 Sep, 2025 1 commit
-
-
vasunvidia authored
* Fused RoPE with combined QKV input. Initial commit for Dropout with 8-bit RNG Fix documentation Initial commit for Fused QKV RoPE WIP Initial tests passing Enable rotary percent and margin Enable CP2, start_positions, interleaved Cleanup test Revert "Fix documentation" This reverts commit 53df10044e7769982bd4af2ae2628e6b7717e715. Revert "Initial commit for Dropout with 8-bit RNG" This reverts commit 301505e24031cbcd679069e1c2cd4d00eedf2dca. Cleanup. Minor cleanup Signed-off-by:
-
- 05 Sep, 2025 1 commit
-
-
Hongbin Liu authored
* fix bugs when enabling --overlap-grad-reduce in mcore Signed-off-by:
-
- 04 Sep, 2025 2 commits
-
-
Casper authored
* fix cross entropy Signed-off-by:
-
Zhongbo Zhu authored
* add noop to comp amax Signed-off-by:
-
- 03 Sep, 2025 1 commit
-
-
Daniel Stokes authored
Signed-off-by:
-
- 02 Sep, 2025 1 commit
-
-
Selvaraj Anandaraj authored
* Create GPU relaod buffers on main stream Signed-off-by:
-
- 31 Aug, 2025 1 commit
-
-
vasunvidia authored
* Add dropout kernel with 8-bit RNG Co-authored-by:
-
- 29 Aug, 2025 3 commits
-
-
buptzyb authored
* support bf16+fp8 model Signed-off-by:
-
Daniel Stokes authored
-
oliver könig authored
* build: pull cached wheels Signed-off-by:
-
- 28 Aug, 2025 3 commits
-
-
Zhongbo Zhu authored
[PyTorch][MOE] Tentative Fix For Replacing from_blob with empty for experts receiving zero tokens (#2134) use torch empty for empty shape instead of from_blob Signed-off-by:
-
Paweł Gadziński authored
* Compute amax in normalization forward in current scaling in untuned kernels Signed-off-by:
-
Charlene Yang authored
* disable determinism for sm100+ and cudnn<9.14 Signed-off-by:
-
- 27 Aug, 2025 1 commit
-
-
Kshitij Lakhani authored
Signed-off-by:
-
- 26 Aug, 2025 1 commit
-
-
Xin Yao authored
expose more activation functions Signed-off-by:
-
- 23 Aug, 2025 1 commit
-
-
Tim Moon authored
* Return dummy wgrad tensors when requested by Mcore Signed-off-by:
Jan Bielak <janekb04@icloud.com> Signed-off-by:
-
- 20 Aug, 2025 3 commits
-
-
Tim Moon authored
Avoid garbage collection when capturing a CUDA Graph Signed-off-by: -
Paweł Gadziński authored
* code drop Signed-off-by:
-
Jan Bielak authored
* Add rmsnorm_bwd_add Signed-off-by:
-
- 18 Aug, 2025 1 commit
-
-
Xin Yao authored
* check if the given recipe is supported in fp8_autocast Signed-off-by:
-
