diff --git a/tools/infer/predict_det.py b/tools/infer/predict_det.py
index 234c41bf03079f6be66bbae5426ec1927661d867..625f87abc39fc0e9d7683f72dafec1d53324873a 100755
--- a/tools/infer/predict_det.py
+++ b/tools/infer/predict_det.py
@@ -42,6 +42,7 @@ class TextDetector(object):
     def __init__(self, args):
         max_side_len = args.det_max_side_len
         self.det_algorithm = args.det_algorithm
+        self.use_zero_copy_run = args.use_zero_copy_run
         preprocess_params = {'max_side_len': max_side_len}
         postprocess_params = {}
         if self.det_algorithm == "DB":
@@ -138,8 +139,12 @@ class TextDetector(object):
             return None, 0
         im = im.copy()
         starttime = time.time()
-        im = fluid.core.PaddleTensor(im)
-        self.predictor.run([im])
+        if self.use_zero_copy_run:
+            self.input_tensor.copy_from_cpu(im)
+            self.predictor.zero_copy_run()
+        else:
+            im = fluid.core.PaddleTensor(im)
+            self.predictor.run([im])
         outputs = []
         for output_tensor in self.output_tensors:
             output = output_tensor.copy_to_cpu()
diff --git a/tools/infer/predict_rec.py b/tools/infer/predict_rec.py
index 922344152dd8a808d212be9bfc0b61254e82fb20..6a379853a4a7d62cbffcbebbf09e2fb3e2207b27 100755
--- a/tools/infer/predict_rec.py
+++ b/tools/infer/predict_rec.py
@@ -40,6 +40,7 @@ class TextRecognizer(object):
         self.character_type = args.rec_char_type
         self.rec_batch_num = args.rec_batch_num
         self.rec_algorithm = args.rec_algorithm
+        self.use_zero_copy_run = args.use_zero_copy_run
         char_ops_params = {
             "character_type": args.rec_char_type,
             "character_dict_path": args.rec_char_dict_path,
@@ -105,8 +106,12 @@ class TextRecognizer(object):
             norm_img_batch = np.concatenate(norm_img_batch)
             norm_img_batch = norm_img_batch.copy()
             starttime = time.time()
-            norm_img_batch = fluid.core.PaddleTensor(norm_img_batch)
-            self.predictor.run([norm_img_batch])
+            if self.use_zero_copy_run:
+                self.input_tensor.copy_from_cpu(norm_img_batch)
+                self.predictor.zero_copy_run()
+            else:
+                norm_img_batch = fluid.core.PaddleTensor(norm_img_batch)
+                self.predictor.run([norm_img_batch])
 
             if self.loss_type == "ctc":
                 rec_idx_batch = self.output_tensors[0].copy_to_cpu()
diff --git a/tools/infer/utility.py b/tools/infer/utility.py
index ec6705eb04eac93cda8a17b8bee310b88226aea9..4dcdfa677d71af43968d2f6acbc2d26495cae5f8 100755
--- a/tools/infer/utility.py
+++ b/tools/infer/utility.py
@@ -71,6 +71,7 @@ def parse_args():
         default="./ppocr/utils/ppocr_keys_v1.txt")
     parser.add_argument("--use_space_char", type=bool, default=True)
     parser.add_argument("--enable_mkldnn", type=bool, default=False)
+    parser.add_argument("--use_zero_copy_run", type=bool, default=False)
     return parser.parse_args()
 
 
@@ -105,8 +106,12 @@ def create_predictor(args, mode):
     #config.enable_memory_optim()
     config.disable_glog_info()
 
-    # use zero copy
-    config.switch_use_feed_fetch_ops(True)
+    if args.use_zero_copy_run:
+        config.delete_pass("conv_transpose_eltwiseadd_bn_fuse_pass")
+        config.switch_use_feed_fetch_ops(False)
+    else:
+        config.switch_use_feed_fetch_ops(True)
+
     predictor = create_paddle_predictor(config)
     input_names = predictor.get_input_names()
     input_tensor = predictor.get_input_tensor(input_names[0])