Skip to content

GitLab

Unverified Commit ebda73dc authored by Qiaolin Yu's avatar Qiaolin Yu Committed by GitHub
Browse files

Use cutlass fp4 gemm by default (#11813)

parent f4f8a1b4
Show whitespace changes
Inline Side-by-side
Showing with 1 addition and 1 deletion
python/sglang/srt/layers/quantization/modelopt_quant.py
View file @ ebda73dc
...@@ -79,7 +79,7 @@ CUTEDSL_MOE_SCALAR_INPUT_SCALE = get_bool_env_var( ...@@ -79,7 +79,7 @@ CUTEDSL_MOE_SCALAR_INPUT_SCALE = get_bool_env_var(
"SGLANG_CUTEDSL_MOE_SCALAR_INPUT_SCALE", "true" "SGLANG_CUTEDSL_MOE_SCALAR_INPUT_SCALE", "true"
) )
USE_CUTLASS_BACKEND_FOR_FP4_GEMM = get_bool_env_var( USE_CUTLASS_BACKEND_FOR_FP4_GEMM = get_bool_env_var(
"SGLANG_USE_CUTLASS_BACKEND_FOR_FP4_GEMM" "SGLANG_USE_CUTLASS_BACKEND_FOR_FP4_GEMM", "true"
) )
# TODO make it true by default when the DeepEP PR is merged # TODO make it true by default when the DeepEP PR is merged
CUTEDSL_MOE_NVFP4_DISPATCH = get_bool_env_var( CUTEDSL_MOE_NVFP4_DISPATCH = get_bool_env_var(
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment