biondizzle d789f5e3e0 Add CCCL include path for CUTLASS 3.x

2026-05-13 23:18:26 +00:00

..

__pycache__

Implement TileLang NVFP4 mega_moe L1/L2 kernels

2026-05-13 22:36:58 +00:00

cutlass_nvfp4_gemm

Add CCCL include path for CUTLASS 3.x

2026-05-13 23:18:26 +00:00

__init__.py

Fix __init__.py: remove private imports

2026-05-13 21:43:47 +00:00

nvfp4_dequant.py

Implement TileLang NVFP4 mega_moe L1/L2 kernels

2026-05-13 22:36:58 +00:00

nvfp4_mega_moe.py

feat: CUTLASS NVFP4 block-scaled GEMM kernel (native SM100 Blackwell)

2026-05-13 23:11:15 +00:00

symm_buffer.py

Initial: TileLang NVFP4 mega_moe kernel package

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

tilelang_kernels.py

Implement TileLang NVFP4 mega_moe L1/L2 kernels

2026-05-13 22:36:58 +00:00

tilelang_nvfp4_gemm.py

Native NVFP4 TileLang kernel: tcgen05 block-scaled MMA

2026-05-13 23:02:06 +00:00

weight_transform.py

Fix: keep scales as float8_e4m3fn, don't pack to uint32 (min_all_cuda unsupported)

2026-05-13 21:54:39 +00:00