cnn+rnn+attention

def attention(inputs, attention_size, time_major=False):
    if isinstance(inputs, tuple):
        # In case of Bi-RNN, concatenate the forward and the backward RNN outputs.
        inputs = tf.concat(inputs, 2)

    if time_major:
        # (T,B,D) => (B,T,D)
        inputs = tf.transpose(inputs, [1, 0, 2])

    inputs_shape = inputs.shape
    sequence_length = inputs_shape[1].value  # the length of sequences processed in the antecedent RNN layer
    hidden_size = inputs_shape[2].value  # hidden size of the RNN layer

    # Attention mechanism
    W_omega = tf.Variable(tf.random_normal([hidden_size, attention_size], stddev=0.1))
    b_omega = tf.Variable(tf.random_normal([attention_size], stddev=0.1))
    u_omega = tf.Variable(tf.random_normal([attention_size], stddev=0.1))

    v = tf.tanh(tf.matmul(tf.reshape(inputs, [-1, hidden_size]), W_omega) + tf.reshape(b_omega, [1, -1]))
    vu = tf.matmul(v, tf.reshape(u_omega, [-1, 1]))
    exps = tf.reshape(tf.exp(vu), [-1, sequence_length])
    alphas = exps / tf.reshape(tf.reduce_sum(exps, 1), [-1, 1])

    # Output of Bi-RNN is reduced with attention vector
    output = tf.reduce_sum(inputs * tf.reshape(alphas, [-1, sequence_length, 1]), 1)

    return output


# cnn的输出为B*4*8*128
chunk_size = 128  
chunk_n = 32  
rnn_size = 256  
attention_size = 50
n_output_layer = MAX_CAPTCHA*CHAR_SET_LEN   # 输出层  

# 定义待训练的神经网络  
def recurrent_neural_network(): 
    data = crack_captcha_cnn()    
   
    data = tf.reshape(data, [-1, chunk_n, chunk_size])  
    data = tf.transpose(data, [1,0,2])  
    data = tf.reshape(data, [-1, chunk_size])  
    data = tf.split(data,chunk_n)
    
    # 只用RNN
    #layer = {'w_':tf.Variable(tf.random_normal([rnn_size, n_output_layer])), 'b_':tf.Variable(tf.random_normal([n_output_layer]))} 
    #lstm_cell = tf.contrib.rnn.BasicLSTMCell(rnn_size) 
    #outputs, status = tf.contrib.rnn.static_rnn(lstm_cell, data, dtype=tf.float32)  
    #ouput = tf.add(tf.matmul(outputs[-1], layer['w_']), layer['b_'])  
    
    # RNN + Attention    
    lstm_cell = tf.contrib.rnn.BasicLSTMCell(rnn_size) 
    outputs, status = tf.contrib.rnn.static_rnn(lstm_cell, data, dtype=tf.float32)     
    attention_output = attention(outputs, attention_size, True)
    
    # output
    drop = tf.nn.dropout(attention_output, keep_prob)
    # Fully connected layer
    W = tf.Variable(tf.truncated_normal([rnn_size, n_output_layer], stddev=0.1), name="W")
    b = tf.Variable(tf.constant(0., shape=[n_output_layer]), name="b")
    
    ouput = tf.nn.xw_plus_b(drop, W, b, name="scores")
    return ouput