removed hard-coded instances of self.latent_dim = 100

gan/gan.py

@@ -34,7 +34,7 @@ class GAN():
         self.generator = self.build_generator()
 
         # The generator takes noise as input and generates imgs
-        z = Input(shape=(100,))
+        z = Input(shape=(self.latent_dim,))
         img = self.generator(z)
 
         # For the combined model we will only train the generator
@@ -112,7 +112,7 @@ class GAN():
             idx = np.random.randint(0, X_train.shape[0], batch_size)
             imgs = X_train[idx]
 
-            noise = np.random.normal(0, 1, (batch_size, 100))
+            noise = np.random.normal(0, 1, (batch_size, self.latent_dim))
 
             # Generate a batch of new images
             gen_imgs = self.generator.predict(noise)
@@ -126,7 +126,7 @@ class GAN():
             #  Train Generator
             # ---------------------
 
-            noise = np.random.normal(0, 1, (batch_size, 100))
+            noise = np.random.normal(0, 1, (batch_size, self.latent_dim))
 
             # Train the generator (to have the discriminator label samples as valid)
             g_loss = self.combined.train_on_batch(noise, valid)
@@ -140,7 +140,7 @@ class GAN():
 
     def sample_images(self, epoch):
         r, c = 5, 5
-        noise = np.random.normal(0, 1, (r * c, 100))
+        noise = np.random.normal(0, 1, (r * c, self.latent_dim))
         gen_imgs = self.generator.predict(noise)
 
         # Rescale images 0 - 1