diff --git a/keras_unet_collection/_model_unet_3plus_2d.py b/keras_unet_collection/_model_unet_3plus_2d.py
index f956a7f..af018ad 100644
--- a/keras_unet_collection/_model_unet_3plus_2d.py
+++ b/keras_unet_collection/_model_unet_3plus_2d.py
@@ -241,10 +241,10 @@ def unet_3plus_2d(input_size, n_labels, filter_num_down, filter_num_skip='auto',
     X_decoder = unet_3plus_2d_backbone(IN, filter_num_down, filter_num_skip, filter_num_aggregate, 
                                        stack_num_down=stack_num_down, stack_num_up=stack_num_up, activation=activation, 
                                        batch_norm=batch_norm, pool=pool, unpool=unpool, name=name)
+    X_decoder = X_decoder[::-1]
+    
     if deep_supervision:
-
         OUT_stack = []
-        X_decoder = X_decoder[::-1]
         L_out = len(X_decoder)
         
         print('----------\ndeep_supervision = True\nnames of output tensors are listed as follows (the last one is the final output):')
@@ -279,7 +279,7 @@ def unet_3plus_2d(input_size, n_labels, filter_num_down, filter_num_skip='auto',
             OUT_stack.append(X)
 
         OUT_stack.append(
-            CONV_output(X_decoder[-1], n_labels, kernel_size=3, 
+            CONV_output(X_decoder[0], n_labels, kernel_size=3, 
                         activation=activation, name='{}_output_final'.format(name)))
         if output_activation:
             print('\t{}_output_final_activation'.format(name))
@@ -289,7 +289,7 @@ def unet_3plus_2d(input_size, n_labels, filter_num_down, filter_num_skip='auto',
         model = Model([IN,], OUT_stack)
 
     else:
-        OUT = CONV_output(X_decoder[-1], n_labels, kernel_size=3, 
+        OUT = CONV_output(X_decoder[0], n_labels, kernel_size=3, 
                           activation=activation, name='{}_output_final'.format(name))
 
         model = Model([IN,], [OUT,])
diff --git a/setup.py b/setup.py
index 6d69156..d2e7d0e 100644
--- a/setup.py
+++ b/setup.py
@@ -5,7 +5,7 @@
 
 setuptools.setup(
     name = "keras-unet-collection",
-    version = "0.0.7",
+    version = "0.0.7beta",
     author = "Yingkai (Kyle) Sha",
     author_email = "yingkaisha@gmail.com",
     description = "The Tensorflow, Keras implementation of U-net, U-net++, R2U-net, Attention U-net, ResUnet-a, U^2-Net, and UNET 3+.",
diff --git a/user_guid.ipynb b/user_guid.ipynb
index 73fdf22..af11a07 100644
--- a/user_guid.ipynb
+++ b/user_guid.ipynb
@@ -283,7 +283,7 @@
     }
    ],
    "source": [
-    "unet3plus = models.unet_3plus_2d((None, None, 3), n_labels=2, filter_num_down=[64, 128, 256, 512], \n",
+    "unet3plus = models.unet_3plus_2d((512, 512, 3), n_labels=2, filter_num_down=[64, 128, 256, 512], \n",
     "                                  filter_num_skip=[64, 64, 64], filter_num_aggregate=256, \n",
     "                                  stack_num_down=2, stack_num_up=1, activation='ReLU', output_activation='Sigmoid',\n",
     "                                  batch_norm=False, pool=True, unpool=False, deep_supervision=True, name='unet3plus')"