Address comments

Signed-off-by: Pavani Majety <pmajety@nvidia.com>

Author: pavanimajety

vllm/model_executor/layers/quantization/compressed_tensors/schemes/compressed_tensors_w4a4_nvfp4.py

@@ -52,7 +52,7 @@ def __init__(self):
             assert has_flashinfer(), f"FlashInfer is required for {self.backend}"
         elif envs.VLLM_NVFP4_GEMM_BACKEND == "cutlass":
             self.backend = "cutlass"
-            assert cutlass_fp4_supported(), "Cutlass is required for {self.backend}"
+            assert cutlass_fp4_supported(), f"Cutlass is required for {self.backend}"
 
         if self.backend == "none":
             raise ValueError(

vllm/model_executor/layers/quantization/modelopt.py

@@ -949,7 +949,7 @@ def __init__(self, quant_config: ModelOptNvFp4Config) -> None:
             assert has_flashinfer(), f"FlashInfer is required for {self.backend}"
         elif envs.VLLM_NVFP4_GEMM_BACKEND == "cutlass":
             self.backend = "cutlass"
-            assert cutlass_fp4_supported(), "Cutlass is required for {self.backend}"
+            assert cutlass_fp4_supported(), f"Cutlass is required for {self.backend}"
 
         if self.backend == "none":
             raise ValueError(