Commits · fab715715993c74fd7be43e6834a782762380891 · OpenDAS / TransformerEngine

You need to sign in or sign up before continuing.

07 Jun, 2025 2 commits

Fix all framework build from PR 1666 (#1857) · fab71571
Kirthi Shankar Sivamani authored Jun 06, 2025
```
Remove extra include
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>
```
fab71571

[PyTorch] Get `skip_fp8_weight_update` only in CUDA Graph Capturing (#1854) · beffb297

Xin Yao authored Jun 07, 2025



only get skip_fp8_weight_update in fp8_graph_capturing
Signed-off-by: Xin Yao <xiny@nvidia.com>
Co-authored-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

beffb297

06 Jun, 2025 3 commits

[Common] Missing CUDA driver deallocations in Userbuffers (#1812) · 05f3b573

Alp Dener authored Jun 06, 2025



* added missing deallocs in Userbuffers destroyer
Signed-off-by: Alp Dener <adener@nvidia.com>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci



---------
Signed-off-by: Alp Dener <adener@nvidia.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

05f3b573

[JAX] GroupedQuantizer and GroupedScaledTensor (#1666) · 7948779c

Phuong Nguyen authored Jun 06, 2025



* refactor the multi_stream utils + implement nvte_multi_tensor_quantize in TE/Common

* implement GroupedQuantizer and grouped_quantize in jaxx

* fix logical_axes_names for transpose tensor in ScaledTensor
Signed-off-by: Phuong Nguyen <phuonguyen@nvidia.com>

---------
Signed-off-by: Phuong Nguyen <phuonguyen@nvidia.com>
Co-authored-by: Hua Huang <huah@nvidia.com>
Co-authored-by: Ming Huang <mingh@nvidia.com>

7948779c

[PyTorch] FP8 Subchannel Recipe With FP8 Gather And Configurable Scaling... · 9985b02c

Zhongbo Zhu authored Jun 05, 2025


[PyTorch] FP8 Subchannel Recipe With FP8 Gather And Configurable Scaling Factor Tensor Swizzling (#1707)

* functional kernel for columnwise + no-transpose option, still hacky
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* pass all quantizer unit tests
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci



* refactor, add gemm ready api
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci



* make format options private members, simplify api
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* swizzle scales right before gemm
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* bug fix of single layer test
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* attempt to fix lint issue
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci



* fp8 gather pass, need minor refine
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci



* fix return_layernorm_output_gathered case
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* remove special cases, add sanity check before gemm
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* fix lint
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* lint ungrouped imports
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci



* Implement dequantize for compact 1D blocks.
Signed-off-by: Keith Wyss <kwyss@nvidia.com>

* add more unit test with dequantize compact supported
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* lint again
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* make ag for subchannel respect async
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* zero tolerance in distributed test
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* fix zero tolerance test
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* resolve rebase issues
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* lint & format
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* fix lint
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* clean up
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* bug fix
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* relax rtol for fp32 distributed test
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* fix some ci issue
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* fix ci test failure in debug mode
Signed-off-by: zhongboz <zhongboz@nvidia.com>

* Force row-wise and column-wise data to have same data format

Prototype "all-gather usage" in quantizer.
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Remove dead logic for high-precision AGs
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci



* Debug FP8 block-wise tensor tests
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Debug distributed test
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Handle case where LayerNormLinear returns gathered norm output
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* fix debug mode
Signed-off-by: zhongboz <zhongboz@nvidia.com>

---------
Signed-off-by: zhongboz <zhongboz@nvidia.com>
Signed-off-by: Keith Wyss <kwyss@nvidia.com>
Signed-off-by: Tim Moon <tmoon@nvidia.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Keith Wyss <kwyss@nvidia.com>
Co-authored-by: Tim Moon <tmoon@nvidia.com>
Co-authored-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>

9985b02c

05 Jun, 2025 4 commits

[JAX] Fix OTYPE for FP8 GEMM (#1838) · 6123d7e0

Phuong Nguyen authored Jun 05, 2025



* fix otype for fp8 gemm
Signed-off-by: Phuong Nguyen <phuonguyen@nvidia.com>
---------
Signed-off-by: Phuong Nguyen <phuonguyen@nvidia.com>

6123d7e0

Use versioned flavor of get driver entrypoint function (#1835) · 557f0cb5

Przemyslaw Tredak authored Jun 05, 2025



* Use versioned flavor of get driver entrypoint function
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

* Update the check to call the versioned API starting with CUDA 12.5 where
it was added
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

* Dynamically find entrypoint functions
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



* Error checking
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

* Lint fix
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

---------
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

557f0cb5

[JAX] Fix 1x quantize kernel availability check on hopper (#1845) · f64d1459

jberchtold-nvidia authored Jun 05, 2025



* Fix 1x quantize kernel availability check on hopper

---------
Signed-off-by: Jeremy Berchtold <jberchtold@nvidia.com>

f64d1459

Fix `NVTE_FRAMEWORK=all` installation (#1850) · 12af02f2

Kirthi Shankar Sivamani authored Jun 04, 2025



* Fix NVTE_FRAMEWORK=all
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* fix
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Fix
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Workflow tests and fixes
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Fix jax install
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Fix
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Update dep
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Add numpy
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Add dep
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

---------
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

12af02f2

04 Jun, 2025 1 commit
- Remove deprecated global option for debug build (#1848) · 97e493f4
  Kirthi Shankar Sivamani authored Jun 04, 2025
```
RM deprecated global option for debug build
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>
```
  97e493f4
03 Jun, 2025 4 commits

[PyTorch] Miscellaneous fixes for attention (#1780) · 151a0af6

Charlene Yang authored Jun 04, 2025



* add missing args in cross-attn
Signed-off-by: Charlene Yang <charleney@nvidia.com>

* allow thd for TELayer
Signed-off-by: Charlene Yang <charleney@nvidia.com>

* add CP note for reordering
Signed-off-by: Charlene Yang <charleney@nvidia.com>

* fix wording about CP
Signed-off-by: Charlene Yang <charleney@nvidia.com>

* add modulo cpx2 requirement
Signed-off-by: Charlene Yang <charleney@nvidia.com>

* add example of token reordering
Signed-off-by: Charlene Yang <charleney@nvidia.com>

* improve the CP docstring
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci



* tweak CP wording
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci



* test thd TELayer
Signed-off-by: Charlene Yang <charleney@nvidia.com>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci



* add enc_dec_kv for decoder
Signed-off-by: Charlene Yang <charleney@nvidia.com>

* fix cross attn in decoder
Signed-off-by: Charlene Yang <charleney@nvidia.com>

* fix unfused + bshd/sbhd + telayer
Signed-off-by: Charlene Yang <charleney@nvidia.com>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci



* remove debugging
Signed-off-by: Charlene Yang <charleney@nvidia.com>

---------
Signed-off-by: Charlene Yang <charleney@nvidia.com>
Signed-off-by: Charlene Yang <8636796+cyanguwa@users.noreply.github.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

151a0af6

Update list of authorized CI users (#1840) · 75fe5601

Vladimir Cherepanov authored Jun 03, 2025



Add to CI users
Signed-off-by: Vladimir Cherepanov <vcherepanov@nvidia.com>
Co-authored-by: Vladimir Cherepanov <vcherepanov@nvidia.com>

75fe5601

Bump cuDNN FE (#1842) · 8b3ba9d6
Kirthi Shankar Sivamani authored Jun 03, 2025
```
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>
```
8b3ba9d6
[JAX] Fix NVTETensor leak in attention.cpp (#1841) · f3d77f6c
jberchtold-nvidia authored Jun 03, 2025
```
Fix NVTETensor leak in attention.cpp
Signed-off-by: Jeremy Berchtold <jberchtold@nvidia.com>
```
f3d77f6c

02 Jun, 2025 2 commits

Minor improvements to runtime error checks during library loading (#1837) · c1417116
Kirthi Shankar Sivamani authored Jun 02, 2025
```
minor build improvements
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>
```
c1417116

[JAX] Use 1x quantization + jax transpose for performance for tensor-scaling (#1830) · 62f5c9ee

jberchtold-nvidia authored Jun 02, 2025



* Use 1x quantization + jax transpose on BW for performance
Signed-off-by: Jeremy Berchtold <jberchtold@nvidia.com>

* Use 1x quantization on Hopper as well as it is also faster
Signed-off-by: Jeremy Berchtold <jberchtold@nvidia.com>

* Undo architecture check helper function
Signed-off-by: Jeremy Berchtold <jberchtold@nvidia.com>

* Lint
Signed-off-by: Jeremy Berchtold <jberchtold@nvidia.com>

---------
Signed-off-by: Jeremy Berchtold <jberchtold@nvidia.com>

62f5c9ee

31 May, 2025 1 commit

[PyTorch][Jax] Add warning for missing SOs if both frameworks are installed (#1834) · c6a9e261

Kirthi Shankar Sivamani authored May 30, 2025



* 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: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

---------
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>
Co-authored-by: Alp Dener <adener@nvidia.com>

c6a9e261

30 May, 2025 2 commits

Quantizer update when recipe was changed (#1814) · d5d78333

Evgeny Tsykunov authored May 30, 2025



* 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: Evgeny <etsykunov@nvidia.com>

* Add test
Signed-off-by: Evgeny <etsykunov@nvidia.com>

* Move _quantizer to the QuantizedTensorBase
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



* Fix import
Signed-off-by: Evgeny Tsykunov <etsykunov@etsykunov-mlt.client.nvidia.com>

---------
Signed-off-by: Evgeny Tsykunov <etsykunov@etsykunov-mlt.client.nvidia.com>
Signed-off-by: Evgeny <etsykunov@nvidia.com>
Co-authored-by: Evgeny Tsykunov <etsykunov@etsykunov-mlt.client.nvidia.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Przemyslaw Tredak <ptredak@nvidia.com>

d5d78333

Avoid searching unnecessary dirs for shared libs (#1801) · 204add8f

Tim Moon authored May 30, 2025


Signed-off-by: Tim Moon <tmoon@nvidia.com>
Signed-off-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>
Co-authored-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

204add8f

29 May, 2025 4 commits

[JAX] Support SWA in CP Ring Attn THD striped sharding (#1810) · 855fa653

Hua Huang authored May 29, 2025



* 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: Hua Huang <huah@nvidia.com>

[pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci



Remove unused check
Signed-off-by: Hua Huang <huah@nvidia.com>

---------
Signed-off-by: Hua Huang <huah@nvidia.com>

855fa653

Avoid memory allocations and deallocations when creating NVTETensor (#1813) · 4292653c

Przemyslaw Tredak authored May 29, 2025



* Changed the Tensor allocation strategy
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

* Fixes
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

* Disable debug flag
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 the double free error
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
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

* Fixed pyTorch recipe extension
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
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

* Fix
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

* Hide TensorAllocator and fix the usage in LayerNorm
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



* Cleaning
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
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

* Fix permutation
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

---------
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

4292653c

[PyTorch] Linear op avoids saving input tensor if weight grad is not needed (#1817) · 41909dc8

Tim Moon authored May 28, 2025



* Linear op avoids saving input tensor if weight grad is not needed
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Linear op forward avoids producing quantized tensors with unnecessary usages
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Fix linter warnings
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Avoid unnecessary usages in fused linear ops
Signed-off-by: Tim Moon <tmoon@nvidia.com>

---------
Signed-off-by: Tim Moon <tmoon@nvidia.com>

41909dc8

[JAX] Removes unneccessary reshapes for FP8 GEMM (#1820) · c9e8e305

Phuong Nguyen authored May 28, 2025



* removes unnecessary reshapes for FP8 GEMM

* use nn.jax.scaled_matmul
Signed-off-by: Phuong Nguyen <phuonguyen@nvidia.com>

---------
Signed-off-by: Phuong Nguyen <phuonguyen@nvidia.com>

c9e8e305

28 May, 2025 3 commits

[JAX] FP8 GEMM via dot_general + direct quant (#1819) · 355c4e42

Phuong Nguyen authored May 28, 2025



* fp8 gemm with direct quant
Signed-off-by: Phuong Nguyen <phuonguyen@nvidia.com>

---------
Signed-off-by: Phuong Nguyen <phuonguyen@nvidia.com>

355c4e42

[JAX] Update jax_scaled_masked_softmax to match TE kernel implementation (#1822) · 4732ed76
jberchtold-nvidia authored May 28, 2025
```
Update jax_scaled_masked_softmax to match TE kernel implementation
Signed-off-by: Jeremy Berchtold <jberchtold@nvidia.com>
```
4732ed76

Fix multi-framework runtime lib loading (#1825) · 30e30811

Kirthi Shankar Sivamani authored May 27, 2025



* Fix single FW build with multi FW available
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Some fixes
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Fixes
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* sug
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

---------
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

30e30811

27 May, 2025 1 commit

Updated README - Added conda installation (#1826) · 9627b073

Santosh Bhavani authored May 27, 2025



* added conda installation
Signed-off-by: Santosh Bhavani <santosh.bhavani@live.com>

* fix for pypi
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

---------
Signed-off-by: Santosh Bhavani <santosh.bhavani@live.com>
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>
Co-authored-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

9627b073

23 May, 2025 2 commits

Fix the failing test cases in the CI (#1806) · cd37379d

Przemyslaw Tredak authored May 23, 2025



* Modify the test cases
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

* Make the tests reproducible on different machines
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

* Fixed the cache of the gamma_in_weight_dtype setting
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

* Reinstate the tests
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



* More verbose code and comments
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>

---------
Signed-off-by: Przemek Tredak <ptredak@nvidia.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

cd37379d

Fix test.sh scripts to test pure-JAX implementations (#1805) · fe9a786c

jberchtold-nvidia authored May 23, 2025



* Fix env variable name in test.sh scripts to properly test pure-JAX implementations
Signed-off-by: Jeremy Berchtold <jberchtold@nvidia.com>

* Update test scripts to use pure-JAX impl in encoder

test_custom_call_compute.py already uses pure-JAX impl as
reference so testing the pure-JAX impl against itself would be
redundant. The encoder tests have their own implementation so
testing the pure-JAX impl of primitives is still useful.
Signed-off-by: Jeremy Berchtold <jberchtold@nvidia.com>

* Update qa/L0_jax_unittest/test.sh
Co-authored-by: Phuong Nguyen <phuonguyen@nvidia.com>
Signed-off-by: jberchtold-nvidia <158520091+jberchtold-nvidia@users.noreply.github.com>

---------
Signed-off-by: Jeremy Berchtold <jberchtold@nvidia.com>
Signed-off-by: jberchtold-nvidia <158520091+jberchtold-nvidia@users.noreply.github.com>
Co-authored-by: Phuong Nguyen <phuonguyen@nvidia.com>

fe9a786c

22 May, 2025 7 commits

[PyTorch] Activation ops support fusing backward pass with quantize (#1804) · e4c051f1
Tim Moon authored May 22, 2025
```
Activation ops support fusing backward pass with quantize
Signed-off-by: Tim Moon <tmoon@nvidia.com>
```
e4c051f1
Add docs for missing FP8 recipes. (#1816) · 1669b3f4
Kirthi Shankar Sivamani authored May 22, 2025
```
Document all recipes
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>
```
1669b3f4
[JAX] Make primitive names more granular for better disabling granularity (#1811) · b17f3f4e
jberchtold-nvidia authored May 22, 2025
```
Make primitive names more granular for better disabling granularity
Signed-off-by: Jeremy Berchtold <jberchtold@nvidia.com>
```
b17f3f4e

Build support for cuda 13 (#1809) · 00328ac7

Kirthi Shankar Sivamani authored May 22, 2025



* Build support for cuda 13
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Fix build for cudnn 8.9*; cuda 12.1
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* readd include
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

---------
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

00328ac7

Remove `comm_gemm_overlap` doc (#1815) · 6262280e
Kirthi Shankar Sivamani authored May 22, 2025
```
Remove comm_gemm_overlap docs
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>
```
6262280e
[JAX] Fix incorrectly skipped test_quantize_dbias tests (#1808) · 0cd1cd8e
jberchtold-nvidia authored May 22, 2025
```
Fix incorrectly skipped test_quantize_dbias tests
Signed-off-by: Jeremy Berchtold <jberchtold@nvidia.com>
```
0cd1cd8e
[PyTorch] Fix saved_tensors access in Ops Fuser (#1807) · 9c436d53
Paweł Gadziński authored May 22, 2025
```
fix saved_tensors
Signed-off-by: Pawel Gadzinski <pgadzinski@nvidia.com>
```
9c436d53

21 May, 2025 2 commits

fix model parallel encoder to be properly sharded params (#1794) · 097afc00

Sudhakar Singh authored May 21, 2025



* fix model parallel encoder to be properly sharded
Signed-off-by: Sudhakar Singh <sudhakars@nvidia.com>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci



---------
Signed-off-by: Sudhakar Singh <sudhakars@nvidia.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

097afc00

Add missing docs for C API (#1803) · cd11e00d

Kirthi Shankar Sivamani authored May 20, 2025



* Add missing docs for C API
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Grammar, typos, copy-paste errors
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* remove contiguous word
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

* Better wording
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

---------
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>

cd11e00d

20 May, 2025 2 commits

[PyTorch] Add docstring for CP load balancing (#1802) · d35afe12
Charlene Yang authored May 21, 2025
```
add docstring for CP
Signed-off-by: Charlene Yang <charleney@nvidia.com>
```
d35afe12

[Pytorch] NVIDIA-DL-Framework-Inspect support – part 4 – documentation (#1611) · 3e50d531

Paweł Gadziński authored May 20, 2025



* docs drop
Signed-off-by: Pawel Gadzinski <pgadzinski@nvidia.com>

* a
Signed-off-by: Pawel Gadzinski <pgadzinski@nvidia.com>

* fix
Signed-off-by: Pawel Gadzinski <pgadzinski@nvidia.com>

* Update docs/debug/1_getting_started.rst
Co-authored-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: Przemyslaw Tredak <ptrendx@gmail.com>
Signed-off-by: Paweł Gadziński <62263673+pggPL@users.noreply.github.com>

* fixes
Signed-off-by: Pawel Gadzinski <pgadzinski@nvidia.com>

* fix imgs
Signed-off-by: Pawel Gadzinski <pgadzinski@nvidia.com>

---------
Signed-off-by: Pawel Gadzinski <pgadzinski@nvidia.com>
Signed-off-by: Paweł Gadziński <62263673+pggPL@users.noreply.github.com>
Co-authored-by: Przemyslaw Tredak <ptrendx@gmail.com>

3e50d531