Pytorch binding for cublas grouped gemm#2669

Open

vthumbe1503 wants to merge 38 commits intoNVIDIA:mainfrom

vthumbe1503:users/vthumbe/pytorch_binding_for_cublas_gemm

Commits on Feb 6, 2026

PyTorch-Python GroupedTensor
ksivaman
committed

Commits on Feb 10, 2026

Commits on Mar 9, 2026

Commits on Mar 10, 2026

Merge branch 'main' into jberchtold/te-common-mxfp8-grouped-gemm-plus-fixes
vthumbe1503
authored
fix the test
vthumbe1503
committed
fix merge conflict
vthumbe1503
committed
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot]
committed
Random padding condition shouldnt be done for mxfp8
vthumbe1503
authored
Remove incorrect comment
vthumbe1503
authored
Merge branch 'main' into jberchtold/te-common-mxfp8-grouped-gemm-plus-fixes
vthumbe1503
authored
CUBLAS > 13.2 is enough
vthumbe1503
authored
merge the PR
vthumbe1503
committed
CUBLAS version needed for MXFP8 indeed seems to be 13.3
vthumbe1503
authored
Merge branch 'jberchtold/te-common-mxfp8-grouped-gemm-plus-fixes' of https://github.com/jberchtold-nvidia/TransformerEngine into users/vthumbe/pytorch_binding_for_cublas_gemm
vthumbe1503
committed
all changes for grouped gemm
vthumbe1503
committed
Merge branch 'users/vthumbe/pytorch_binding_for_cublas_gemm' of github.com:vthumbe1503/TransformerEngine into users/vthumbe/pytorch_binding_for_cublas_gemm
vthumbe1503
committed
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot]
committed

Commits on Mar 11, 2026