- 17 Oct, 2025 2 commits
-
-
Neil Tenenholtz authored
* Fix test of FSDP2 by correcting init logic and applying autocast This fixes multiple issues in the FSDP2 test, namely 1. Previously fp8 init was performed when `args.fp8_init == False`. I have updated the logic to match what I presume was intended by leveraging the nullcontext context manager. 2. `te.fp8_autocast` was previously not called; the recipe was created but was unused. The autocast context manager now wraps the model's computation. Signed-off-by:
Neil Tenenholtz <ntenenz@users.noreply.github.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Fix typo Signed-off-by:
Tim Moon <4406448+timmoon10@users.noreply.github.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Fix bug when constructing context for model init Signed-off-by:
pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by:
-
Kshitij Lakhani authored
* Fix imports in test for deprecated jax.experimental.pjit Signed-off-by:
Kshitij Lakhani <klakhani@nvidia.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Fix: Pass NamedSharding instead of PartitionSpec to compare_ops() so that when the in and out sharding is used to create a jitted function, it has the mesh info Signed-off-by:
Kshitij Janardan Lakhani <klakhani@login-eos01.eos.clusters.nvidia.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci --------- Signed-off-by:
-
- 14 Oct, 2025 3 commits
-
-
Kirthi Shankar Sivamani authored
* Initial API change Signed-off-by:
Kirthi Shankar Sivamani <ksivamani@nvidia.com> * Change all imports and api Signed-off-by:
-
Kshitij Lakhani authored
* Add BRCM support when creating a test mask for fused attn 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 1 commit
-
-
Peter St. John authored
Signed-off-by:
Peter St. John <pstjohn@nvidia.com> Co-authored-by:
-
- 09 Oct, 2025 1 commit
-
-
jberchtold-nvidia authored
Signed-off-by:
Jeremy Berchtold <jberchtold@nvidia.com> Co-authored-by:
Phuong Nguyen <phuonguyen@nvidia.com>
-
- 08 Oct, 2025 1 commit
-
-
Hua Huang authored
* Try async copy of grouped GEMM group_sizes data Signed-off-by:
Hua Huang <huah@nvidia.com> --------- Signed-off-by:
-
- 07 Oct, 2025 1 commit
-
-
vcherepanov-nv authored
Signed-off-by:Vladimir Cherepanov <vcherepanov@nvidia.com>
-
- 06 Oct, 2025 1 commit
-
-
Phuong Nguyen authored
* not fuse bias for output all reduction case + unit tests Signed-off-by:
-
- 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:
-
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 1 commit
-
-
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 1 commit
-
-
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:
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:
jberchtold-nvidia <158520091+jberchtold-nvidia@users.noreply.github.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:
Xin Yao <xiny@nvidia.com> * Remove exceptions from destructors 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 2 commits
-
-
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:
Przemek Tredak <ptredak@nvidia.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Fix linter warning Signed-off-by:
-
jberchtold-nvidia authored
Address tolerance check for current scaling dact Signed-off-by:
-
- 23 Sep, 2025 1 commit
-
-
Ming-Xu Huang authored
* Adding Amax Primitive and related args. 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
-
-
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 1 commit
-
-
vthumbe1503 authored
* Test working as I think it should work 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 2 commits
-
-
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:
-
vcherepanov-nv authored
* Extract cpp distributed tests into a separate project Signed-off-by:
-
- 08 Sep, 2025 2 commits
-
-
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:
-
Ming-Xu Huang authored
* Fixing few issues with multi-process launching. Signed-off-by:
-
- 05 Sep, 2025 2 commits
-
-
jberchtold-nvidia authored
* Custom call tests passing Signed-off-by:
-
vcherepanov-nv authored
* Fix CUDA version in setup.py Signed-off-by:
-
- 04 Sep, 2025 1 commit
-
-
Casper authored
* fix cross entropy Signed-off-by:
-
- 03 Sep, 2025 2 commits
-
-
Kshitij Lakhani authored
* Fix failing tests for dropout=0.1 and bias for fused attn for blackwell Signed-off-by:
-
Daniel Stokes authored
Signed-off-by:
-
- 31 Aug, 2025 1 commit
-
-
vasunvidia authored
* Add dropout kernel with 8-bit RNG Co-authored-by:
-
- 29 Aug, 2025 1 commit
-
-
Daniel Stokes authored
-
- 28 Aug, 2025 3 commits
-
-
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:
-
vcherepanov-nv authored
Signed-off-by:
-
