- 31 May, 2025 1 commit
-
-
Kirthi Shankar Sivamani authored
* Add warning for multi framework case Signed-off-by:
Kirthi Shankar Sivamani <ksivamani@nvidia.com> Co-authored-by:
Alp Dener <adener@nvidia.com> * fix Signed-off-by:
-
- 30 May, 2025 2 commits
-
-
Evgeny Tsykunov authored
* Quantizer update Signed-off-by:
Evgeny Tsykunov <etsykunov@etsykunov-mlt.client.nvidia.com> * Update import Signed-off-by:
Evgeny <etsykunov@nvidia.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Introduce _update_weight_quantizers and _get_weight_tensors/_get_weight_quantizers Signed-off-by:
pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by:
Przemyslaw Tredak <ptredak@nvidia.com>
-
Tim Moon authored
Signed-off-by:
Tim Moon <tmoon@nvidia.com> Signed-off-by:
Tim Moon <4406448+timmoon10@users.noreply.github.com> Co-authored-by:
-
- 29 May, 2025 4 commits
-
-
Hua Huang authored
* Support SWA in CP Ring Attn THD striped sharding Signed-off-by:
Hua Huang <huah@nvidia.com> * Add some comments; move check to _FusedAttnCPWithP2PHelper.check_supported() Signed-off-by:
-
Przemyslaw Tredak authored
* Changed the Tensor allocation strategy Signed-off-by:
-
Tim Moon authored
* Linear op avoids saving input tensor if weight grad is not needed Signed-off-by:
-
Phuong Nguyen authored
* removes unnecessary reshapes for FP8 GEMM * use nn.jax.scaled_matmul Signed-off-by:
Phuong Nguyen <phuonguyen@nvidia.com> --------- Signed-off-by:
-
- 28 May, 2025 3 commits
-
-
Phuong Nguyen authored
* fp8 gemm with direct quant Signed-off-by:
-
jberchtold-nvidia authored
Update jax_scaled_masked_softmax to match TE kernel implementation Signed-off-by:Jeremy Berchtold <jberchtold@nvidia.com>
-
Kirthi Shankar Sivamani authored
* Fix single FW build with multi FW available Signed-off-by:
-
- 23 May, 2025 1 commit
-
-
Przemyslaw Tredak authored
* Modify the test cases Signed-off-by:
-
- 22 May, 2025 5 commits
-
-
Tim Moon authored
Activation ops support fusing backward pass with quantize Signed-off-by: -
Kirthi Shankar Sivamani authored
Document all recipes Signed-off-by: -
jberchtold-nvidia authored
Make primitive names more granular for better disabling granularity Signed-off-by: -
Kirthi Shankar Sivamani authored
* Build support for cuda 13 Signed-off-by:
-
Paweł Gadziński authored
fix saved_tensors Signed-off-by:Pawel Gadzinski <pgadzinski@nvidia.com>
-
- 21 May, 2025 1 commit
-
-
Kirthi Shankar Sivamani authored
* Add missing docs for C API Signed-off-by:
-
- 20 May, 2025 4 commits
-
-
Charlene Yang authored
add docstring for CP Signed-off-by:Charlene Yang <charleney@nvidia.com>
-
Paweł Gadziński authored
* docs drop Signed-off-by:
Przemyslaw Tredak <ptrendx@gmail.com> Signed-off-by:
Paweł Gadziński <62263673+pggPL@users.noreply.github.com> * Update docs/debug/1_getting_started.rst Co-authored-by:
-
Peter St. John authored
* Use an empty torch tensor to indicate no fp8 information in extra_state Signed-off-by:
Peter St. John <pstjohn@nvidia.com> * Add huggingface from_pretrained / save_pretrained tests Adds integration tests to ensure models containing TransformerLayer objects can be saved and loaded using the from_pretrained and save_pretrained methods. Signed-off-by:
-
guyueh1 authored
* Fix split_overlap_rs aggregate=True chunk offset calculation Signed-off-by:
Guyue Huang <guyueh@nvidia.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Add unit test for aggregate=True Signed-off-by:
-
- 19 May, 2025 2 commits
-
-
Evgeny Tsykunov authored
* Check tensor-recipe compatibility Signed-off-by:
-
Paweł Gadziński authored
* tests drop Signed-off-by:
-
- 16 May, 2025 2 commits
-
-
Selvaraj Anandaraj authored
* Added token ignoring for CE loss Signed-off-by:
Selvaraj Anandaraj <selvaraja@cw-dfw-cs-001-login-01.cm.cluster> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Added tests Signed-off-by:
root <root@cw-dfw-h100-004-210-013.cm.cluster> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci --------- Signed-off-by:
-
jberchtold-nvidia authored
* [JAX] Update flax module param initialization to support logical partitioning axes Signed-off-by:
-
- 15 May, 2025 1 commit
-
-
Kirthi Shankar Sivamani authored
* Cleanup runtime library loading Signed-off-by:
-
- 14 May, 2025 2 commits
-
-
Peter St. John authored
Signed-off-by:
-
Kirthi Shankar Sivamani authored
* rm unused swizzle extensions Signed-off-by:
-
- 13 May, 2025 3 commits
-
-
Evgeny Tsykunov authored
* Set sequence_parallel before super().__init__() in norm modules Signed-off-by:
-
Charlene Yang authored
* disable sm89 and cuDNN < 9.11 for KV caching Signed-off-by:
-
Kirthi Shankar Sivamani authored
* Disallow kwargs for pybind extensions and release GIL if possible Signed-off-by:
-
- 12 May, 2025 1 commit
-
-
jberchtold-nvidia authored
This reverts commit 5bee81e2 . Signed-off-by:
-
- 11 May, 2025 1 commit
-
-
Kirthi Shankar Sivamani authored
* First pass refactor Signed-off-by:
-
- 09 May, 2025 1 commit
-
-
Tim Moon authored
* Avoid spurious warning with non-FP8 GroupedLinear Signed-off-by:
-
- 08 May, 2025 3 commits
-
-
Paweł Gadziński authored
* features drop Signed-off-by:
-
Li Tao authored
* use lru to cache torch.Tensor() Signed-off-by:
lit <lit@nvidia.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * remove duplicated definition Signed-off-by:
-
Xiaowei Ren authored
Signed-off-by:
Xiaowei Ren <xren@nvidia.com> Co-authored-by:
-
- 07 May, 2025 2 commits
-
-
Tim Moon authored
* Initial work toward restoring UB support in te.Sequential Signed-off-by:
-
Peter St. John authored
Signed-off-by:
-
- 06 May, 2025 1 commit
-
-
Przemyslaw Tredak authored
* Changes to Linear Signed-off-by:
-
