[Bugfix][CI] Fix Marlin FP8 Linear Kernel for Compressed Tensors Format (#38092)

Signed-off-by: BadrBasowid <Badr.Basowid@gmail.com>
Signed-off-by: BadrBasowid <61441185+BadrBasowid@users.noreply.github.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

vllm/model_executor/layers/quantization/compressed_tensors/schemes/compressed_tensors_w8a8_fp8.py

@@ -188,6 +188,9 @@ class CompressedTensorsW8A8Fp8(CompressedTensorsScheme):
         if self.strategy == QuantizationStrategy.BLOCK:
             maybe_post_process_fp8_weight_block(layer)
 
+        if hasattr(self, "fp8_linear"):
+            self.fp8_linear.process_weights_after_loading(layer)
+
     def apply_weights(
         self,
         layer: torch.nn.Module,

vllm/model_executor/layers/quantization/modelopt.py

@@ -705,6 +705,9 @@ class ModelOptFp8PbWoLinearMethod(LinearMethodBase):
 
         layer.weight_scale = Parameter(scale.contiguous(), requires_grad=False)
 
+        if hasattr(self, "fp8_linear"):
+            self.fp8_linear.process_weights_after_loading(layer)
+
     def apply(
         self,
         layer: torch.nn.Module,