snake achieve cifar as well as relu

code/resnet18/resnet18_snake.py

@@ -33,7 +33,8 @@ def snake(x):
 ## Chargement et normalisation des données
 resnet18 = tf.keras.datasets.cifar10
 (train_images, train_labels), (test_images, test_labels) = resnet18.load_data()
-
+train_images = train_images.astype('float32')
+test_images = test_images.astype('float32')
 
 from sklearn.model_selection import train_test_split
 train_images, val_images, train_labels, val_labels = train_test_split(train_images,train_labels, test_size = 0.2,shuffle = True)
@@ -98,8 +99,8 @@ model.compile(sgd, loss='categorical_crossentropy', metrics=['accuracy'])
 
 history = model.fit(train_images,
          train_labels,
-         batch_size=64,
-         epochs=4,
+         batch_size=256,
+         epochs=50,
          validation_data=(val_images, val_labels),
          )