tesnsorflow 使用LSTM进行分类的例程

import tensorflow as tf
import sys
from tensorflow.examples.tutorials.mnist import input_data

# this is data
mnist = input_data.read_data_sets('MNIST_data', one_hot=True)

# hyperparameters
lr = 0.001
training_iters = 100000
batch_size = 128

n_inputs = 28   # MNIST data input (img shape: 28*28)
n_steps = 28    # time steps
n_hidden_units = 128   # neurons in hidden layer
n_classes = 10      # MNIST classes (0-9 digits)

# tf Graph input
x = tf.placeholder(tf.float32, [None, n_steps, n_inputs])
y = tf.placeholder(tf.float32, [None, n_classes])

# Define weights
weights = {
    # (28, 128)
    'in': tf.Variable(tf.random_normal([n_inputs, n_hidden_units])),
    # (128, 10)
    'out': tf.Variable(tf.random_normal([n_hidden_units, n_classes]))
}
biases = {
    # (128, )
    'in': tf.Variable(tf.constant(0.1, shape=[n_hidden_units, ])),
    # (10, )
    'out': tf.Variable(tf.constant(0.1, shape=[n_classes, ]))
}


def RNN(X, weights, biases):
    # hidden layer for input to cell
    ########################################
    #X(128 batch,28 steps,28 inputs)
    #==>(128*28,28 inputs)
    X = tf.reshape(X,[-1,n_inputs])    
    #==>(128 batch*28 steps,128 hidden)
    X_in = tf.matmul(X,weights['in'])+biases['in']
    #==>(128 batch,28 steps,128 hidden)
    X_in = tf.reshape(X_in,[-1,n_steps,n_hidden_units])
    # cell
    ##########################################
    #same to define active function
    lstm_cell = tf.nn.rnn_cell.BasicLSTMCell(n_hidden_units,forget_bias=1.0,state_is_tuple=True)
    #lstm cell is divided into two parts(c_state,m_state)
    _init_state = lstm_cell.zero_state(batch_size,dtype=tf.float32)
    
    #choose rnn how to work,lstm just is one kind of rnn,use lstm_cell for active function,set initial_state
    outputs,states = tf.nn.dynamic_rnn(lstm_cell,X_in,initial_state=_init_state,time_major=False)   
        
    # hidden layer for output as the final results
    #############################################
    results = tf.matmul(states[1],weights['out']) + biases['out']   
    
    #unpack to list [(batch,outputs)]*steps
    #outputs = tf.unpack(tf.transpose(outputs,[1,0,2])) # state is the last outputs
    #results = tf.matmul(outputs[-1],weights['out']) + biases['out']
    return results


pred = RNN(x, weights, biases)
cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(pred, y))
train_op = tf.train.AdamOptimizer(lr).minimize(cost)

correct_pred = tf.equal(tf.argmax(pred, 1), tf.argmax(y, 1))
accuracy = tf.reduce_mean(tf.cast(correct_pred, tf.float32))

init = tf.initialize_all_variables()
with tf.Session() as sess:
    sess.run(init)
    step = 0
    while step * batch_size < training_iters:
        batch_xs, batch_ys = mnist.train.next_batch(batch_size)
        batch_xs = batch_xs.reshape([batch_size, n_steps, n_inputs])
        sess.run([train_op], feed_dict={
            x: batch_xs,
            y: batch_ys,
        })
        if step % 20 == 0:
            print(sess.run(accuracy, feed_dict={
            x: batch_xs,
            y: batch_ys,
        }))
        step += 1

说明：

1、该LSTM模型的结构是28*128*10

def RNN(X, weights, biases):
    # hidden layer for input to cell
    ########################################
    #X(128 batch,28 steps,28 inputs)
    #==>(128*28,28 inputs)
    X = tf.reshape(X,[-1,n_inputs])
    #==>(128 batch*28 steps,128 hidden)
    X_in = tf.matmul(X,weights['in'])+biases['in']
    #==>(128 batch,28 steps,128 hidden)
    X_in = tf.reshape(X_in,[-1,n_steps,n_hidden_units])
    # cell
    ##########################################
    #same to define active function
    lstm_cell = tf.nn.rnn_cell.BasicLSTMCell(n_hidden_units,forget_bias=1.0,state_is_tuple=True)
    #lstm cell is divided into two parts(c_state,m_state)
    _init_state = lstm_cell.zero_state(batch_size,dtype=tf.float32)

    #choose rnn how to work,lstm just is one kind of rnn,use lstm_cell for active function,set initial_state
    outputs,states = tf.nn.dynamic_rnn(lstm_cell,X_in,initial_state=_init_state,time_major=False)

    # hidden layer for output as the final results
    #############################################
    results = tf.matmul(states[1],weights['out']) + biases['out']

    #unpack to list [(batch,outputs)]*steps
    #outputs = tf.unpack(tf.transpose(outputs,[1,0,2])) # state is the last outputs
    #results = tf.matmul(outputs[-1],weights['out']) + biases['out']
    return results

讲解如下：

1、

    #X(128 batch,28 steps,28 inputs)
    #==>(128*28,28 inputs)
    X = tf.reshape(X,[-1,n_inputs])
    #==>(128 batch*28 steps,128 hidden)
    X_in = tf.matmul(X,weights['in'])+biases['in']
    #==>(128 batch,28 steps,128 hidden)
    X_in = tf.reshape(X_in,[-1,n_steps,n_hidden_units])

说明：首先,对于输入我们先进行加权,因为在LSTM单元中,忘记门,输入门等各个部件都需要输入的加权和,直接就在这边做好

2、

    # cell
    ##########################################
    #same to define active function
    lstm_cell = tf.nn.rnn_cell.BasicLSTMCell(n_hidden_units,forget_bias=1.0,state_is_tuple=True)
    #lstm cell is divided into two parts(c_state,m_state)
    _init_state = lstm_cell.zero_state(batch_size,dtype=tf.float32)

说明：在这边定义我们使用的LSTM单元,其实定义LSTM单元跟定义激活函数类似LSTM单元其实就是替换掉sigmoid激活函数,LSTM单元仅仅是一个细胞单元.然后LSTM细胞的输出是一个tuple,tuple = (状态值,激活值)

3、

    #choose rnn how to work,lstm just is one kind of rnn,use lstm_cell for active function,set initial_state
    outputs,states = tf.nn.dynamic_rnn(lstm_cell,X_in,initial_state=_init_state,time_major=False)

说明：定义使用某种LSTM单元的RNN网络

4、

    # hidden layer for output as the final results
    #############################################
    results = tf.matmul(states[1],weights['out']) + biases['out']

说明：最终的输出使用的是最后的step,隐层的激活值再和输出层的权重进行加权

5、

init = tf.initialize_all_variables()
with tf.Session() as sess:
    sess.run(init)
    step = 0
    while step * batch_size < training_iters:
        batch_xs, batch_ys = mnist.train.next_batch(batch_size)
        batch_xs = batch_xs.reshape([batch_size, n_steps, n_inputs])
        sess.run([train_op], feed_dict={
            x: batch_xs,
            y: batch_ys,
        })
        if step % 20 == 0:
            print(sess.run(accuracy, feed_dict={
            x: batch_xs,
            y: batch_ys,
        }))
        step += 1

说明：这是训练的过程

6、使用LSTM单元的RNN也有前向过程和反向传播过程,只是前向过程是在时序上进行前向传播。所以,使用LSTM单元的RNN必须要记录每个step,LSTM单元的激活值和状态。