- 01 Oct, 2025 1 commit
-
-
Evgeny Tsykunov authored
* Introduce QuantizerBase Signed-off-by:
Evgeny <etsykunov@nvidia.com> * Expose as a first-class API Signed-off-by:
Tim Moon <4406448+timmoon10@users.noreply.github.com> * Merge main followup Signed-off-by:
pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by:
-
- 30 Sep, 2025 3 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:
Przemyslaw Tredak <ptrendx@gmail.com> 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:
Jan Bielak <jbielak@nvidia.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * fix Signed-off-by:
Pawel Gadzinski <pgadzinski@nvidia.com> * fix Signed-off-by:
zhongboz <zhongboz@nvidia.com> Co-authored-by:
Kirthi Shankar Sivamani <ksivamani@nvidia.com> Signed-off-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:
Xin Yao <xiny@nvidia.com> * Remove exceptions from destructors Signed-off-by:
Tim Moon <tmoon@nvidia.com> * fix weired dispatch in ln/rmsnorm 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:
Przemek Tredak <ptredak@nvidia.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci --------- 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:
-
- 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 2 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:
-
alan yang authored
* feat: add cutlass group gemm support Signed-off-by:
-
- 17 Sep, 2025 2 commits
-
-
Daniel Stokes authored
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:
-
- 09 Sep, 2025 1 commit
-
-
Autumn1998 authored
* add bf16/fp32 token-per-expert on the moe-loss-computation on router fusion 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
-
-
vcherepanov-nv authored
* Fix CUDA version in setup.py Signed-off-by:
-
- 04 Sep, 2025 1 commit
-
-
Zhongbo Zhu authored
* add noop to comp amax Signed-off-by:
-
- 31 Aug, 2025 1 commit
-
-
vasunvidia authored
* Add dropout kernel with 8-bit RNG Co-authored-by:
-
- 29 Aug, 2025 2 commits
-
-
Xin Yao authored
* add checks to cuda kernel launch and cuda API calls Signed-off-by:
-
Daniel Stokes authored
-
- 26 Aug, 2025 4 commits
-
-
Vladimir Cherepanov authored
* Pick up cuBLASMp during build Signed-off-by:
-
jberchtold-nvidia authored
Revert "[Common] PDL for Blockwise Quantization (#2066)" This reverts commit ebca6153 . Signed-off-by:
-
vcherepanov-nv authored
* Bump cuDNN FE to 1.14.0 Signed-off-by:
-
jberchtold-nvidia authored
Revert "[Common] PDL for Quantization Kernels (#2001)" This reverts commit bfab8c67 . Signed-off-by:
-
- 20 Aug, 2025 2 commits
-
-
Tim Moon authored
* Fix incorrect version checks for atomic GEMM Signed-off-by:
-
Jan Bielak authored
* Add rmsnorm_bwd_add Signed-off-by:
-
- 15 Aug, 2025 1 commit
-
-
Jan Bielak authored
* Add `nvte_cublas_gemm_scaled` Signed-off-by:
-
- 14 Aug, 2025 3 commits
-
-
Kirthi Shankar Sivamani authored
Add launch bounds to swizzle kernel, use empty scale inv Signed-off-by: -
Xin Yao authored
* reduce driver calls Signed-off-by:
-
Kshiteej K authored
* fix: update grad_output quant to avoid redundant work Signed-off-by:
kshitij12345 <kshitijkalambarkar@gmail.com> * add test Signed-off-by:
-
- 13 Aug, 2025 1 commit
-
-
Xin Yao authored
* enable PDL for blockwise qunatization kernels Signed-off-by:
-
- 12 Aug, 2025 1 commit
-
-
Jan Bielak authored
* Compute amax in normalization forward in current scaling in untuned kernels Signed-off-by:
-
- 09 Aug, 2025 1 commit
-
-
Daniel Stokes authored
* fix: Add stream synchronization before destroying MPI communicator (#1979) Signed-off-by:
-
- 07 Aug, 2025 1 commit
-
-
Xin Yao authored
* PDL for MXFP8 Quantize Signed-off-by:
-
- 06 Aug, 2025 1 commit
-
-
Xin Yao authored
* for loop Signed-off-by:
-
- 04 Aug, 2025 1 commit
-
-
Tim Moon authored
* Add basic kernel for swapping first two tensor dims Signed-off-by:
-
- 01 Aug, 2025 2 commits
-
-
Shang Zhang authored
Fix tensor numel() return dtype The original dytpe int would be an issue if the tensor element numbers are larger than int32's range (which is not a super large number in real workloads) Signed-off-by:Shang Zhang <shangz@nvidia.com>
-
Autumn1998 authored
* fix bug if all values<0 Signed-off-by:
-
- 31 Jul, 2025 1 commit
-
-
Jan Bielak authored
* Compute amax in normalization kernels as long as the pointer is provided, even if using non quantized output Signed-off-by:
-
- 29 Jul, 2025 1 commit
-
-
Jan Bielak authored
* Compute amax in activation kernels when the output pointer is provided, even for non-fp8 outputs Signed-off-by:
-
