Fix quark fp8 format loading. (ROCm#395)

* fix quark fp8 loading

* fix undefined variables

---------

Co-authored-by: Bowen Bao <[email protected]>

Author: fxmarty-amd

vllm/model_executor/layers/quantization/quark/schemes/quark_w8a8_fp8.py

@@ -35,23 +35,26 @@ def process_weights_after_loading(self, layer) -> None:
         # tensor scales (thus N scales being passed to the kernel),
         # requantize so we can always run per tensor
         if self.qscheme == "per_tensor":
-            max_w_scale, weight = requantize_with_max_scale(
-                weight=layer.weight,
-                weight_scale=layer.weight_scale,
-                logical_widths=layer.logical_widths,
-            )
-
             if current_platform.is_rocm():
                 weight, max_w_scale, input_scale = normalize_e4m3fn_to_e4m3fnuz(
-                    weight=weight,
-                    weight_scale=max_w_scale,
+                    weight=layer.weight,
+                    weight_scale=layer.weight_scale,
                     input_scale=layer.input_scale)
-                if input_scale is not None:
-                    layer.input_scale = Parameter(input_scale,
-                                                  requires_grad=False)
+            else:
+                max_w_scale = layer.weight_scale
+                weight = layer.weight
+                input_scale = layer.input_scape
+
+            max_w_scale, weight = requantize_with_max_scale(
+                weight=weight,
+                weight_scale=max_w_scale,
+                logical_widths=layer.logical_widths,
+            )
 
             layer.weight = Parameter(weight.t(), requires_grad=False)
             layer.weight_scale = Parameter(max_w_scale, requires_grad=False)
+            if input_scale is not None:
+                layer.input_scale = Parameter(input_scale, requires_grad=False)
 
         # If channelwise, scales are already lined up, so just transpose.
         elif self.qscheme == "per_channel":