- 07 Oct, 2025 1 commit
-
-
Phuong Nguyen authored
* reuse amax for current scaling Signed-off-by:
Phuong Nguyen <phuonguyen@nvidia.com> --------- Signed-off-by:
-
- 06 Oct, 2025 1 commit
-
-
Phuong Nguyen authored
* not fuse bias for output all reduction case + unit tests Signed-off-by:
-
- 04 Oct, 2025 2 commits
-
-
Kirthi Shankar Sivamani authored
* Fix in FP8 attention selection logic Signed-off-by:
Kirthi Shankar Sivamani <ksivamani@nvidia.com> * Improve logic Signed-off-by:
-
Tim Moon authored
Signed-off-by:Tim Moon <tmoon@nvidia.com>
-
- 03 Oct, 2025 2 commits
-
-
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:
-
vthumbe1503 authored
Signed-off-by:Varun Thumbe <vthumbe@nvidia.com> *Jax integration for clamped swiglu. This is the continuation of PR which added Clamped Swiglu(used in GPT OSS) support in TE along with Pytorch integration. This PR hooks up the clamped swiglu and dswiglu's nvte APIs to TE Jax.
-
- 02 Oct, 2025 4 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:
-
jberchtold-nvidia authored
Fix shard map issue Signed-off-by:
Jeremy Berchtold <jberchtold@nvidia.com> Co-authored-by:
-
jberchtold-nvidia authored
Fix code block in fp8_autocast docstring Signed-off-by:
-
Tim Moon authored
* Make sure to set usages for linear op quantizers before forward Signed-off-by:
-
- 01 Oct, 2025 7 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:
-
Kirthi Shankar Sivamani authored
Fix passing args to nvfp4 recipe Signed-off-by: -
Kirthi Shankar Sivamani authored
Disable debug build for cutlass GEMM Signed-off-by: -
Evgeny Tsykunov authored
Fix QuantizedTensorBase -> QuantizedTensorStorage Signed-off-by:Evgeny <etsykunov@nvidia.com>
-
Phuong Nguyen authored
fix rng_state shape Signed-off-by:
jberchtold-nvidia <158520091+jberchtold-nvidia@users.noreply.github.com>
-
Phuong Nguyen authored
* rm using_global_amax_of_x Signed-off-by:
-
Evgeny Tsykunov authored
* Introduce QuantizerBase Signed-off-by:
-
- 30 Sep, 2025 4 commits
-
-
jberchtold-nvidia authored
Load modules during initialize Signed-off-by:
JAX Toolbox <jax@nvidia.com>
-
vthumbe1503 authored
* Test working as I think it should work 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:
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:
-
Przemyslaw Tredak authored
* Fix the segfault in the nvfp4 quantization 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:
-
- 27 Sep, 2025 1 commit
-
-
Phuong Nguyen authored
* init cgemm + unit tests * UB bootstrap with NCCL, no MPI dependency * add NVLINK-P2P check + error message * skip tests if no NVLINK available * use std::vector to store ncclComm_t * update misuse of TP warning Signed-off-by:
-
- 23 Sep, 2025 2 commits
-
-
Phuong Nguyen authored
* Rework shardy rules * WAR for compound factor=1 Signed-off-by:
-
Ming-Xu Huang authored
* Adding Amax Primitive and related args. Signed-off-by:
-
- 22 Sep, 2025 2 commits
-
-
Charlene Yang authored
* first draft; debug plan failure Signed-off-by:
Chen Cui <chcui@nvidia.com> Signed-off-by:
-
Phuong Nguyen authored
* remove import jax.extend.ffi 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 4 commits
-
-
Daniel Stokes authored
Signed-off-by: -
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 2 commits
-
-
Daniel Stokes authored
Signed-off-by: -
Jan Bielak authored
* Do not use norm fwd + amax fusion if cudnn backend is requested Signed-off-by:
-
- 15 Sep, 2025 3 commits
-
-
Tim Moon authored
* Add CPU offloading logic to ops. Fix test to compute dgrad. Signed-off-by:
-
Ming-Xu Huang authored
* Applying the original precision as Norm outputs' and activation compuations. Signed-off-by:
-
Paweł Gadziński authored
* fix underflows log issue 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:
-
