[Compressed Tensors] Always clone output for compile robustness (#26849)

Signed-off-by: Kyle Sayers <kylesayrs@gmail.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>

vllm/model_executor/layers/quantization/compressed_tensors/transform/linear.py

@@ -163,7 +163,7 @@ class CompressedTensorsLinearTransformMethod(LinearMethodBase):
         if self.output_transform is not None:
             for part_id, (start, length) in enumerate(self.partition_ranges):
                 x[:, start : start + length] = self.output_transform(
-                    x[:, start : start + length].contiguous(), part_id=part_id
+                    x[:, start : start + length].clone(), part_id=part_id
                 )
 
         return x