Normalize labels

1 files changed, 6 insertions, 3 deletions

diff --git a/ncdcgan.py b/ncdcgan.py
index ccb99d3..ce6b907 100755
--- a/ncdcgan.py
+++ b/ncdcgan.py
@@ -157,6 +157,7 @@ class nCDCGAN():
         X_train = (X_train.astype(np.float32) - 127.5) / 127.5
         X_train = np.expand_dims(X_train, axis=3)
         y_train = y_train.reshape(-1, 1)
+        y_train = (y_train.astype(np.float32)-4.5)/4.5
 
         # Adversarial ground truths
         valid = np.ones((batch_size, 1))
@@ -214,6 +215,7 @@ class nCDCGAN():
         r, c = 2, 5
         noise = np.random.normal(0, 1, (r * c, 100))
         sampled_labels = np.arange(0, 10).reshape(-1, 1)
+        sampled_labels = (sampled_labels.astype(np.float32)-4.5)/4.5
 
         #using dummy_labels would just print zeros to help identify image quality
         #dummy_labels = np.zeros(32).reshape(-1, 1)
@@ -239,6 +241,7 @@ class nCDCGAN():
       noise_test = np.random.normal(0, 1, (10000, 100))
       noise_val = np.random.normal(0, 1, (5000, 100))
 
+      ((labels_val.astype(np.float32)-4.5)/4.5)
       labels_train = np.zeros(55000).reshape(-1, 1)
       labels_test = np.zeros(10000).reshape(-1, 1)
       labels_val = np.zeros(5000).reshape(-1, 1)
@@ -248,9 +251,9 @@ class nCDCGAN():
         labels_test[i*1000:-1] = i
         labels_val[i*500:-1] = i
 
-      train_data = self.generator.predict([noise_train, labels_train])
-      test_data = self.generator.predict([noise_test, labels_test])
-      val_data = self.generator.predict([noise_val, labels_val])
+      train_data = self.generator.predict([noise_train, ((labels_train.astype(np.float32)-4.5)/4.5)])
+      test_data = self.generator.predict([noise_test, ((labels_test.astype(np.float32)-4.5)/4.5)])
+      val_data = self.generator.predict([noise_val,((labels_val.astype(np.float32)-4.5)/4.5)])
 
       labels_train = keras.utils.to_categorical(labels_train, 10)
       labels_test = keras.utils.to_categorical(labels_test, 10)