diff --git a/pix2pix/pix2pix.py b/pix2pix/pix2pix.py
index 2e8652a341..257f056c08 100644
--- a/pix2pix/pix2pix.py
+++ b/pix2pix/pix2pix.py
@@ -29,7 +29,6 @@ def __init__(self):
         self.data_loader = DataLoader(dataset_name=self.dataset_name,
                                       img_res=(self.img_rows, self.img_cols))
 
-
         # Calculate output shape of D (PatchGAN)
         patch = int(self.img_rows / 2**4)
         self.disc_patch = (patch, patch, 1)
@@ -56,18 +55,17 @@ def __init__(self):
 
         # Input images and their conditioning images
         img_A = Input(shape=self.img_shape)
-        img_B = Input(shape=self.img_shape)
 
         # By conditioning on B generate a fake version of A
-        fake_A = self.generator(img_B)
+        fake_A = self.generator(img_A)
 
         # For the combined model we will only train the generator
         self.discriminator.trainable = False
 
         # Discriminators determines validity of translated images / condition pairs
-        valid = self.discriminator([fake_A, img_B])
+        valid = self.discriminator([fake_A, img_A])
 
-        self.combined = Model(inputs=[img_A, img_B], outputs=[valid, fake_A])
+        self.combined = Model(inputs=img_A, outputs=[valid, fake_A])
         self.combined.compile(loss=['mse', 'mae'],
                               loss_weights=[1, 100],
                               optimizer=optimizer)
@@ -171,7 +169,7 @@ def train(self, epochs, batch_size=1, sample_interval=50):
                 # -----------------
 
                 # Train the generators
-                g_loss = self.combined.train_on_batch([imgs_A, imgs_B], [valid, imgs_A])
+                g_loss = self.combined.train_on_batch(imgs_B, [valid, imgs_A])
 
                 elapsed_time = datetime.datetime.now() - start_time
                 # Plot the progress