CNN simples com AUC

# -*- coding: utf-8 -*-
"""
Created on Sat Apr 29 15:54:45 2017

@author: www.deeplearningbrasil.com.br
"""

# Rede Neural convolucional simples para o problema de reconhecimento de dígitos (MNIST)
import tensorflow as tf
import random
# import matplotlib.pyplot as plt

from tensorflow.examples.tutorials.mnist import input_data

tf.set_random_seed(777)  # reproducibility

mnist = input_data.read_data_sets("MNIST_data/", one_hot=True)
# Verifique o site https://www.tensorflow.org/get_started/mnist/beginners para
# mais informações sobre o conjunto de dados

# Parâmetros de aprendizagem
taxa_aprendizado = 0.001
quantidade_maxima_epocas = 10
batch_size = 100

# entrada dos place holders
X = tf.placeholder(tf.float32, [None, 784])
X_img = tf.reshape(X, [-1, 28, 28, 1])   # imagem 28x28x1 (preto e branca)
Y = tf.placeholder(tf.float32, [None, 10])

# L1 ImgIn shape=(?, 28, 28, 1)
W1 = tf.Variable(tf.random_normal([3, 3, 1, 32], stddev=0.01))
#    Conv     -> (?, 28, 28, 32)
#    Pool     -> (?, 14, 14, 32)
L1 = tf.nn.conv2d(X_img, W1, strides=[1, 1, 1, 1], padding='SAME')
L1 = tf.nn.relu(L1)
L1 = tf.nn.avg_pool(L1, ksize=[1, 2, 2, 1],
                    strides=[1, 2, 2, 1], padding='SAME')
'''
Tensor("Conv2D:0", shape=(?, 28, 28, 32), dtype=float32)
Tensor("Relu:0", shape=(?, 28, 28, 32), dtype=float32)
Tensor("MaxPool:0", shape=(?, 14, 14, 32), dtype=float32)
'''

# L2 ImgIn shape=(?, 14, 14, 32)
W2 = tf.Variable(tf.random_normal([3, 3, 32, 64], stddev=0.01))
#    Conv      ->(?, 14, 14, 64)
#    Pool      ->(?, 7, 7, 64)
L2 = tf.nn.conv2d(L1, W2, strides=[1, 1, 1, 1], padding='SAME')
L2 = tf.nn.relu(L2)
L2 = tf.nn.avg_pool(L2, ksize=[1, 2, 2, 1],
                    strides=[1, 2, 2, 1], padding='SAME')
L2_flat = tf.reshape(L2, [-1, 7 * 7 * 64])
'''
Tensor("Conv2D_1:0", shape=(?, 14, 14, 64), dtype=float32)
Tensor("Relu_1:0", shape=(?, 14, 14, 64), dtype=float32)
Tensor("MaxPool_1:0", shape=(?, 7, 7, 64), dtype=float32)
Tensor("Reshape_1:0", shape=(?, 3136), dtype=float32)
'''

# Classificador - Camada Fully Connected entrada 7x7x64 -> 10 saídas
W3 = tf.get_variable("W3", shape=[7 * 7 * 64, 10],
                     initializer=tf.contrib.layers.xavier_initializer())
b = tf.Variable(tf.random_normal([10]))
logits = tf.matmul(L2_flat, W3) + b

# Define a função de custo e o método de otimização
cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=logits, labels=Y))
optimizer = tf.train.AdamOptimizer(taxa_aprendizado).minimize(cost)

# inicializa
sess = tf.Session()
sess.run(tf.global_variables_initializer())

# treina a rede
print('Rede inicialiada. Treinamento inicializado. Tome um cafe...')
for epoca in range(quantidade_maxima_epocas):
    custo_medio = 0
    total_batch = int(mnist.train.num_examples / batch_size)

    for i in range(total_batch):
        batch_xs, batch_ys = mnist.train.next_batch(batch_size)
        feed_dict = {X: batch_xs, Y: batch_ys}
        c, _ = sess.run([cost, optimizer], feed_dict=feed_dict)
        custo_medio += c / total_batch

    print('Epoca:', '%04d' % (epoca + 1), 'perda =', '{:.9f}'.format(custo_medio))

print('Treinamento finalizado!')

# Teste o modelo e verifica a taxa de acerto
correct_prediction = tf.equal(tf.argmax(logits, 1), tf.argmax(Y, 1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))

#Saída da rede
d = tf.cast(tf.argmax(logits, 1),tf.float32)
#Respostas Corretas
y = tf.cast(tf.argmax(Y,1),tf.float32)

#Calcula a área da curva ROC (AUC)
auc, update_auc = tf.contrib.metrics.streaming_auc(d,y)

sess.run(tf.local_variables_initializer())

#Avalia Acurácia
print('Taxa de acerto:', sess.run(accuracy, feed_dict={
      X: mnist.test.images, Y: mnist.test.labels}))

#Avalia AUC
print('Area da curva ROC:', sess.run(update_auc, feed_dict={
      X: mnist.test.images, Y: mnist.test.labels}))


'''
#Se for necessário "quebrar" o teste em batchs substitua os dois prints acima pelo código abaixo
total_batch = int(mnist.test.num_examples / batch_size)
acc = 0
for i in range(total_batch):
    batch_xs, batch_ys = mnist.test.next_batch(batch_size)
    feed_dict = {X: batch_xs, Y: batch_ys}

    #Acurácia parcial
    acc += sess.run(accuracy, feed_dict)/total_batch
    #Atualiza curva ROC
    roc = sess.run(update_auc, feed_dict)

#Resultado final da AUC
roc = sess.run(auc, feed_dict)

#Avalia Acurácia
print('Taxa de acerto:', acc)

#Avalia AUC
print('Area da curva ROC:', roc)

'''