# 生成一个训练batch
def get_next_batch(batch_size=128):
batch_x = np.zeros([batch_size, IMAGE_HEIGHT * IMAGE_WIDTH])#128,60*160
batch_y = np.zeros([batch_size, MAX_CAPTCHA * CHAR_SET_LEN])#128,4*63
# 有时生成图像大小不是(60, 160, 3)
def wrap_gen_captcha_text_and_image():
while True:
text, image = gen_captcha_text_and_image_new()
if image.shape == (60, 160, 3):
return text, image
for i in range(batch_size):
text, image = wrap_gen_captcha_text_and_image()
image = convert2gray(image)
#把image展成一维,然后除以255 eg:[0.94901961 0.96470588 0.96470588...]
batch_x[i, :] = image.flatten() / 255 # (image.flatten()-128)/128 mean为0
batch_y[i, :] = text2vec(text)
return batch_x, batch_y
####################################################################
X = tf.placeholder(tf.float32, [None, IMAGE_HEIGHT * IMAGE_WIDTH])
Y = tf.placeholder(tf.float32, [None, MAX_CAPTCHA * CHAR_SET_LEN])
keep_prob = tf.placeholder(tf.float32) # dropout
# 定义CNN
def crack_captcha_cnn(w_alpha=0.01, b_alpha=0.1):
x = tf.reshape(X, shape=[-1, IMAGE_HEIGHT, IMAGE_WIDTH, 1])
# 3 conv layer
w_c1 = tf.Variable(w_alpha * tf.random_normal([3, 3, 1, 32]))
b_c1 = tf.Variable(b_alpha * tf.random_normal([32]))
conv1 = tf.nn.relu(tf.nn.bias_add(tf.nn.conv2d(x, w_c1, strides=[1, 1, 1, 1], padding='SAME'), b_c1))
conv1 = tf.nn.max_pool(conv1, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
conv1 = tf.nn.dropout(conv1, keep_prob)
w_c2 = tf.Variable(w_alpha * tf.random_normal([3, 3, 32, 64]))
b_c2 = tf.Variable(b_alpha * tf.random_normal([64]))
conv2 = tf.nn.relu(tf.nn.bias_add(tf.nn.conv2d(conv1, w_c2, strides=[1, 1, 1, 1], padding='SAME'), b_c2))
conv2 = tf.nn.max_pool(conv2, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
conv2 = tf.nn.dropout(conv2, keep_prob)
w_c3 = tf.Variable(w_alpha * tf.random_normal([3, 3, 64, 64]))
b_c3 = tf.Variable(b_alpha * tf.random_normal([64]))
conv3 = tf.nn.relu(tf.nn.bias_add(tf.nn.conv2d(conv2, w_c3, strides=[1, 1, 1, 1], padding='SAME'), b_c3))
conv3 = tf.nn.max_pool(conv3, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
conv3 = tf.nn.dropout(conv3, keep_prob)
#60*160*1->30*80*32->15*40*64->8*20*64
# Fully connected layer
w_d = tf.Variable(w_alpha * tf.random_normal([8 * 20 * 64, 1024]))
b_d = tf.Variable(b_alpha * tf.random_normal([1024]))
#dense = tf.reshape(conv3, [-1, w_d.get_shape().as_list()[0]])
#上面一行代码意思与下一行代码一样,10240=8*20*64
dense = tf.reshape(conv3, [-1, 10240])
dense = tf.nn.relu(tf.add(tf.matmul(dense, w_d), b_d))
dense = tf.nn.dropout(dense, keep_prob)
w_out = tf.Variable(w_alpha * tf.random_normal([1024, MAX_CAPTCHA * CHAR_SET_LEN]))
b_out = tf.Variable(b_alpha * tf.random_normal([MAX_CAPTCHA * CHAR_SET_LEN]))
out = tf.add(tf.matmul(dense, w_out), b_out)
# out = tf.nn.softmax(out)
return out
# 训练
def train_crack_captcha_cnn():
output = crack_captcha_cnn()
# loss
# loss = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(output, Y))
loss = tf.reduce_mean(tf.nn.sigmoid_cross_entropy_with_logits(logits=output, labels=Y))
# 最后一层用来分类的softmax和sigmoid有什么不同?
# optimizer 为了加快训练 learning_rate应该开始大,然后慢慢衰
optimizer = tf.train.AdamOptimizer(learning_rate=0.001).minimize(loss)
predict = tf.reshape(output, [-1, MAX_CAPTCHA, CHAR_SET_LEN])
max_idx_p = tf.argmax(predict, 2)
max_idx_l = tf.argmax(tf.reshape(Y, [-1, MAX_CAPTCHA, CHAR_SET_LEN]), 2)
correct_pred = tf.equal(max_idx_p, max_idx_l)
accuracy = tf.reduce_mean(tf.cast(correct_pred, tf.float32))
saver = tf.train.Saver()
with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
step = 0
while True:
batch_x, batch_y = get_next_batch(64)
_, loss_ = sess.run([optimizer, loss], feed_dict={X: batch_x, Y: batch_y, keep_prob: 0.75})
#print("No. %d steps,the loss is %f " %(step, loss_))
# 每100 step计算一次准确率
if step % 100 == 0:
batch_x_test, batch_y_test = get_next_batch(100)
acc = sess.run(accuracy, feed_dict={X: batch_x_test, Y: batch_y_test, keep_prob: 1.})
print("No. %d steps,the acc is %f " %(step, acc))
# 如果准确率大于50%,保存模型,完成训练
if acc > 0.99:
saver.save(sess, "./tmp/ckpt/model", global_step=step)
break
step += 1
然后我们调用train_crack_captcha_cnn()函数进行训练
train_crack_captcha_cnn()