Refactor NVFP4 Linear utils for ModelOpt and CT (#33201)

Signed-off-by: mgoin <mgoin64@gmail.com>
2026-01-30 19:37:42 -05:00
parent 2b465570e6
commit 67ebaff528
12 changed files with 462 additions and 483 deletions
--- a/tests/kernels/moe/modular_kernel_tools/mk_objects.py
+++ b/tests/kernels/moe/modular_kernel_tools/mk_objects.py
@@ -25,7 +25,7 @@ from vllm.model_executor.layers.fused_moe.prepare_finalize import (
 from vllm.model_executor.layers.fused_moe.triton_deep_gemm_moe import (
    TritonOrDeepGemmExperts,
 )
-from vllm.model_executor.layers.quantization.utils.quant_utils import (
+from vllm.model_executor.layers.quantization.utils.nvfp4_utils import (
    cutlass_fp4_supported,
 )
 from vllm.model_executor.layers.quantization.utils.w8a8_utils import (