biondizzle 5d49849156 Fix: torch.zeros doesn't support Float4_e2m1fn_x2 dtype ...

Allocate as uint8 then view as float4_e2m1fn_x2 for padding buffer.

2026-05-19 01:15:24 +00:00

..

kernels/linear/nvfp4

Fix: torch.zeros doesn't support Float4_e2m1fn_x2 dtype

2026-05-19 01:15:24 +00:00

patches

Register CuTeDSL as proper NvFp4LinearKernel for NVFP4 linear layers

2026-05-19 00:44:44 +00:00

cutedsl_quant_method.py

Fix: add abstract create_weights to CuTeDSLNvfp4LinearMethod

2026-05-18 20:40:48 +00:00

nvfp4_cutedsl.py

Fix torch.compile: use custom autograd Function instead of @torch.compiler.disable

2026-05-18 21:38:28 +00:00