biondizzle 6326222d68 Fix: add abstract create_weights to CuTeDSLNvfp4LinearMethod

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

..

patches

Wire CuTeDSL kernels into vLLM: replace all BF16 dequant with native NVFP4

2026-05-18 20:27:42 +00:00

cutedsl_quant_method.py

Fix: add abstract create_weights to CuTeDSLNvfp4LinearMethod

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

nvfp4_cutedsl.py

Wire CuTeDSL kernels into vLLM: replace all BF16 dequant with native NVFP4

2026-05-18 20:27:42 +00:00