biondizzle 80495c0cd6 docs: clarify SF layout remap is in CUDA, not sf_layout.py ...

sf_layout.py was a no-op (return sf) but the actual remap happens
in remap_sf_to_cutlass_kernel in cutlass_nvfp4_gemm.cu. Updated
sf_layout.py to pure reference docs so nobody gets confused again.

2026-05-14 13:04:31 +00:00

..

cutlass_nvfp4_gemm

docs: clarify SF layout remap is in CUDA, not sf_layout.py

2026-05-14 13:04:31 +00:00

__init__.py

cleanup: remove abandoned TileLang and Mojo files

2026-05-14 12:44:47 +00:00

nvfp4_mega_moe.py

fix: rewrite stage_activation with proper E2M1 quantization

2026-05-14 13:02:10 +00:00

symm_buffer.py

Initial: TileLang NVFP4 mega_moe kernel package

2026-05-13 15:44:51 +00:00

weight_transform.py

cleanup: remove abandoned TileLang and Mojo files

2026-05-14 12:44:47 +00:00