[Enhancement] move copy_to_cpu before counting time in infer.py (#2595)

Author: wuyefeilin

Matting/deploy/python/infer.py

@@ -397,11 +397,11 @@ def run(self, imgs, trimaps=None, imgs_dir=None):
 
             self.predictor.run()
 
+            results = output_handle.copy_to_cpu()
+
             if args.benchmark:
                 self.autolog.times.stamp()
 
-            results = output_handle.copy_to_cpu()
-
             results = results.squeeze(1)
             for j in range(args.batch_size):
                 trimap = trimap_inputs[j] if trimaps is not None else None