#!/usr/bin/env python
# coding: utf-8

# In[1]:


from keras.models import Sequential, Model
from keras.datasets import cifar10
from keras.layers import Dense, Activation, Flatten, Input, merge, Convolution2D, MaxPooling2D
from keras.utils import np_utils
from keras.callbacks import EarlyStopping
from keras.utils.visualize_util import model_to_dot, plot
import numpy as np
from IPython.display import SVG

from matplotlib import pyplot as plt
get_ipython().run_line_magic('matplotlib', 'inline')
plt.style.use("ggplot")
np.random.seed(13)


# In[2]:


batch_size = 256
nb_classes = 10
nb_epoch = 20
nb_filter = 10

img_rows, img_cols = 32, 32
img_channels = 3

(X_train, y_train), (X_test, y_test) = cifar10.load_data()
print('X_train shape:', X_train.shape)
print(X_train.shape[0], 'train samples')
print(X_test.shape[0], 'test samples')
X_train = X_train.astype('float32')
X_test = X_test.astype('float32')
X_train /= 255
X_test /= 255

Y_train = np_utils.to_categorical(y_train, nb_classes)
Y_test = np_utils.to_categorical(y_test, nb_classes)


# In[3]:


model = Sequential()
model.add(Convolution2D(nb_filter, 3, 3, input_shape=(img_channels, img_rows, img_cols), border_mode="same", activation="relu"))
model.add(Convolution2D(nb_filter, 3, 3, border_mode="same", activation="relu"))
model.add(Convolution2D(nb_filter, 3, 3, border_mode="same", activation="relu"))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Convolution2D(nb_filter, 3, 3, border_mode="same", activation="relu"))
model.add(Convolution2D(nb_filter, 3, 3, border_mode="same", activation="relu"))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Flatten())
model.add(Dense(512, activation="relu"))
model.add(Dense(nb_classes, activation="softmax"))

model.compile(loss='categorical_crossentropy',
              optimizer='adadelta',
              metrics=['accuracy'])

SVG(model_to_dot(model, show_shapes=True).create(prog='dot', format='svg'))


# In[4]:


in_img = Input(shape=(img_channels, img_rows, img_cols))
x = Convolution2D(nb_filter, 3, 3, border_mode="same", activation="relu")(in_img)
for _ in range(2):
    y = Convolution2D(nb_filter, 3, 3, border_mode="same", activation="relu")(x)
    y = Convolution2D(nb_filter, 3, 3, border_mode="same")(y)
    x = merge([x, y], mode="sum")
    x = Activation("relu")(x)
    x = MaxPooling2D(pool_size=(2, 2))(x)

x = Flatten()(x)
x = Dense(512, activation="relu")(x)
x = Dense(nb_classes, activation="softmax")(x)


residual = Model(input=in_img, output=x)

residual.compile(loss='categorical_crossentropy',
              optimizer='adadelta',
              metrics=['accuracy'])

SVG(model_to_dot(residual, show_shapes=True).create(prog='dot', format='svg'))


# In[ ]:


cnn = model.fit(X_train, Y_train, batch_size=batch_size, nb_epoch=nb_epoch, validation_data=(X_test, Y_test))


# In[ ]:


resi = residual.fit(X_train, Y_train, batch_size=batch_size, nb_epoch=nb_epoch, validation_data=(X_test, Y_test))


# In[ ]:


x = range(nb_epoch)
plt.plot(x, cnn.history['acc'], label="cnn train")
plt.plot(x, cnn.history['val_acc'], label="cnn val")
plt.plot(x, resi.history['acc'], label="resi train")
plt.plot(x, resi.history['val_acc'], label="resi val")
plt.title("accuracy")
plt.legend(loc='center left', bbox_to_anchor=(1, 0.5))
plt.show()

plt.plot(x, cnn.history['loss'], label="cnn train")
plt.plot(x, cnn.history['val_loss'], label="cnn val")
plt.plot(x, resi.history['loss'], label="resi train")
plt.plot(x, resi.history['val_loss'], label="resi val")
plt.title("loss")
plt.legend(loc='center left', bbox_to_anchor=(1, 0.5))
plt.show()


# In[ ]: