diff --git a/train.py b/train.py
index cc09307cf136948fcac65c7ef4d6308ef4a94ab3..6441d1db12534b980266eb1a90f31020aff86766 100644
--- a/train.py
+++ b/train.py
@@ -106,16 +106,17 @@ if __name__ == "__main__":
     #------------------------------------------------------#
     model = YoloBody(anchors_mask, num_classes)
     weights_init(model)
-    #------------------------------------------------------#
-    #   权值文件请看README，百度网盘下载
-    #------------------------------------------------------#
-    print('Load weights {}.'.format(model_path))
-    device          = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
-    model_dict      = model.state_dict()
-    pretrained_dict = torch.load(model_path, map_location = device)
-    pretrained_dict = {k: v for k, v in pretrained_dict.items() if np.shape(model_dict[k]) == np.shape(v)}
-    model_dict.update(pretrained_dict)
-    model.load_state_dict(model_dict)
+    if model_path != '':
+        #------------------------------------------------------#
+        #   权值文件请看README，百度网盘下载
+        #------------------------------------------------------#
+        print('Load weights {}.'.format(model_path))
+        device          = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+        model_dict      = model.state_dict()
+        pretrained_dict = torch.load(model_path, map_location = device)
+        pretrained_dict = {k: v for k, v in pretrained_dict.items() if np.shape(model_dict[k]) == np.shape(v)}
+        model_dict.update(pretrained_dict)
+        model.load_state_dict(model_dict)
 
     model_train = model.train()
     if Cuda: