CLDNN

论文

《CONVOLUTIONAL, LONG SHORT-TERM MEMORY,FULLY CONNECTED DEEP NEURAL NETWORKS》

CLDNN网络结构描述

考虑到LSTM局限性，CLDNN通过结合三种网络结构，来解决问题：

• 把特征输入到CNN层，降低谱差异性；
• 把CNN的输出输入到LSTM建模时序特征；
• 把LSTM的输出作为DNN的输入，减少LSTM隐层的变化，使得特征转化到更可分的空间。

Tensorflow网络实现测试Mnist

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

tf.reset_default_graph() 

def cnn(input_tensor):
    input_tensor = tf.reshape(input_tensor, [-1, 28, 28, 1])
    with tf.name_scope('conv-1') as scope:
        conv1_weights = tf.Variable(initial_value=tf.truncated_normal(shape=[5, 5, 1, 64], dtype=tf.float32, stddev=0.01),
            name='weights'
        )
        conv1_bias = tf.Variable(
            initial_value=tf.truncated_normal(shape=[64], dtype=tf.float32, stddev=0.01),
            name='bias'
        )
        conv1 = tf.nn.conv2d(input_tensor, conv1_weights, strides=[1, 1, 1, 1], padding='SAME')
        conv1 = tf.nn.bias_add(conv1, conv1_bias)
        relu1 = tf.nn.relu(conv1, name=scope)
    with tf.name_scope('pool-1'):
        pool1 = tf.nn.max_pool(relu1, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='VALID')
    with tf.name_scope('conv-2') as scope:
        conv2_weights = tf.Variable(
            initial_value=tf.truncated_normal(shape=[5, 5, 64, 64], dtype=tf.float32, stddev=0.01),
            name='wieghts'
        )
        conv2_bias = tf.Variable(
            initial_value=tf.truncated_normal(shape=[64], dtype=tf.float32, stddev=0.01),
            name='bias'
        )
        conv2 = tf.nn.conv2d(pool1, conv2_weights, strides=[1, 1, 1, 1], padding='SAME')
        conv2 = tf.nn.bias_add(conv2, conv2_bias)
        relu2 = tf.nn.relu(conv2, name=scope)
    return relu2

    
def linear(input_tensor):
    input_shape = input_tensor.get_shape().as_list()
    width = input_shape[1]
    height = input_shape[2]
    maps = input_shape[3]
    flatten_size = width * height * maps
    input_tensor = tf.reshape(input_tensor, shape=[-1, flatten_size])
    
    with tf.name_scope('Linear'):
        dense_weights = tf.Variable(tf.truncated_normal(shape=[flatten_size, 256],stddev=0.1), dtype=tf.float32)
        dense_bias = tf.Variable(tf.truncated_normal(shape=[256]), dtype=tf.float32)
        result = tf.nn.bias_add(tf.matmul(input_tensor, dense_weights), dense_bias)
    return result

# parameters init
l_r = 0.001
training_iters = 100000
batch_size = 128

n_inputs = 16
n_steps = 16
n_hidden_units = 128
n_classes = 10  
  
def lstm(input_tensor):
    with tf.variable_scope('lstm1'):
            lstm1_weights_in = tf.get_variable("weight_in", [n_inputs,n_hidden_units],initializer = tf.random_normal_initializer())
            lstm1_biases_in = tf.get_variable("bias_in", [n_hidden_units,],initializer = tf.constant_initializer(0.1))
            lstm1_weights_out = tf.get_variable("weight_out", [n_hidden_units,n_classes],initializer = tf.random_normal_initializer())
            lstm1_biases_out = tf.get_variable("bias_out", [n_classes,],initializer = tf.constant_initializer(0.1))
            
            input_tensor = tf.reshape(input_tensor, [-1, n_inputs])
            x_in = tf.matmul(input_tensor, lstm1_weights_in) + lstm1_biases_in
            x_in = tf.reshape(x_in, [-1, n_steps, n_hidden_units])
            
            lstm_cell = tf.contrib.rnn.BasicLSTMCell(n_hidden_units, forget_bias=1.0, state_is_tuple=True)
            _init_state = lstm_cell.zero_state(batch_size, dtype=tf.float32)
            outputs,states = tf.nn.dynamic_rnn(lstm_cell, x_in, initial_state=_init_state, time_major=False)
    
            outputs = tf.unstack(tf.transpose(outputs, [1,0,2]))
            results = tf.matmul(outputs[-1], lstm1_weights_out) + lstm1_biases_out
    return results
                            
    
    
def inference(input_tensor):
    net = cnn(input_tensor)
    net = linear(net)
    net = lstm(net)
    return net
    

#load mnist data
mnist = input_data.read_data_sets("../../../datasets/MNIST_data", one_hot=True)

#define placeholder for input
x = tf.placeholder(tf.float32, [None, 28*28*1])
y_ = tf.placeholder(tf.float32, [None, n_classes])

y = inference(x)
cost = tf.reduce_mean(tf.nn.sparse_softmax_cross_entropy_with_logits(labels=tf.argmax(y_, 1), logits=y))


train_op = tf.train.AdamOptimizer(l_r).minimize(cost)

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

#init session
sess = tf.Session()
#init all variables
sess.run(tf.global_variables_initializer())
#start training

#for i in range(training_iters):
for i in range(training_iters):
    #get batch to learn easily
    batch_x, batch_y = mnist.train.next_batch(batch_size)
#    batch_x = batch_x.reshape([batch_size, n_steps, n_inputs])
    sess.run(train_op,feed_dict={x: batch_x, y_: batch_y})
    if i % 50 == 0:
        print(sess.run(accuracy,feed_dict={x: batch_x, y_: batch_y,}))


#test_data = mnist.test.images.reshape([-1, n_steps, n_inputs])
#test_label = mnist.test.labels
#print("Testing Accuracy: ", sess.run(accuracy, feed_dict={x: test_data, y: test_label}))

参考：

https://github.com/xxoospring/CLDNNs