CNN做序列標注問題（tensorflow）

一、搭建簡單的CNN做序列標注代碼

import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt
    
  
TIME_STEPS = 15# backpropagation through time 的time_steps
BATCH_SIZE = 1#50
INPUT_SIZE = 1 # x數據輸入size
LR = 0.05  # learning rate
num_tags = 2 
# 定義一個生成數據的 get_batch function:
def get_batch():
    xs = np.array([[[[2], [3], [4], [5], [5], [5], [1], [5], [3], [2], [5], [5], [5], [3], [5]]]])
    res = np.array([[0, 0, 1, 1, 1, 1, 0, 1, 0, 0, 1, 1, 1, 0, 1]])
    ys = np.zeros([1,TIME_STEPS,2])
    for i in range(TIME_STEPS):
        if(res[0,i] == 0):
            ys[0,i,0] = 1
            ys[0,i,1] = 0
        else:
            ys[0,i,0] = 0
            ys[0,i,1] = 1
        
    return [xs, res,ys]
    
# 定義 CNN 的主體結構
class CNN(object):
    def __init__(self, n_steps, input_size, num_tags, batch_size):
        self.n_steps = n_steps
        self.input_size = input_size
        self.num_tags = num_tags
        self.batch_size = batch_size
        #卷積神將網絡的輸入：[batch, in_height, in_width, in_channels]，在自然語言處理中height為1
        self.xs = tf.placeholder(tf.float32, [self.batch_size,1, self.n_steps, self.input_size], name='xs')
        #做序列標注，第二維對應好輸入的n_steps，相當於每個時刻的輸入都有一個輸出
        self.ys = tf.placeholder(tf.int32, [self.batch_size, self.n_steps,self.num_tags], name='ys')#
        
        self.featureNum = 10#提取10個特征
        
        #[卷積核的高度，卷積核的寬度，圖像通道數，卷積核個數]
        W_conv1 = self.weight_variable([1,3,1,self.featureNum])#提取10個特征
        #對應10個卷積核輸出
        b_conv1 = self.bias_varibale([self.featureNum]) 
    
        #卷積操作
        layer_conv1  = tf.nn.conv2d(self.xs, W_conv1,strides=[1, 1, 1, 1],padding="SAME",) + b_conv1
        #激勵層
        layer_conv1  = tf.nn.relu(layer_conv1)
        #最大值池化  本處去除池化層為了后續計算簡便
        #layer_pool1  = tf.nn.max_pool(layer_conv1,
        #                              [1, 1, 3, 1],[1,1,1,1],padding='VALID') 
        layer_pool1 = layer_conv1

        # 全連接層  映射到self.n_steps x self.num_tags
        layer_pool1 = tf.reshape(layer_pool1,[self.n_steps,self.featureNum])
        W_fc1  = self.weight_variable([self.featureNum,self.num_tags])
        b_fc1  = self.bias_varibale([self.num_tags])
        h_fc1  = tf.matmul(layer_pool1, W_fc1) + b_fc1
        #激勵層
        h_fc1 = tf.nn.relu(h_fc1)
        #softmax 歸一化
        self.y_conv = tf.nn.softmax(h_fc1)        
        self.label = tf.reshape(self.ys,[self.n_steps,2])
        self.cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(labels=self.label, logits=self.y_conv))
        #梯度下降
        self.train_op = tf.train.AdamOptimizer(LR).minimize(self.cost)      
        self.pred = tf.argmax(self.y_conv,axis = 1)
       
    def weight_variable(self,shape):
        initial=tf.truncated_normal(shape, mean=0.0, stddev=0.1)
        return tf.Variable(initial)
    def bias_varibale(self,shape):
        initial=tf.constant(0,1,shape=shape)
        return tf.Variable(initial)   
    
# 訓練CNN
if __name__ == '__main__':
       
    # 搭建 CNN 模型
    model = CNN(TIME_STEPS, INPUT_SIZE, num_tags, BATCH_SIZE)
    sess = tf.Session()
    sess.run(tf.global_variables_initializer())
      
    # matplotlib可視化
    plt.ion()  # 設置連續 plot
    plt.show()  
    # 訓練多次
    for i in range(150):
        xs, res,ys = get_batch()  # 提取 batch data
        # 初始化 data
        feed_dict = {
            model.xs: xs,
            model.ys: ys,
        }        
        # 訓練
        _, cost,pred = sess.run(
            [model.train_op, model.cost,  model.pred],
            feed_dict=feed_dict)

    
        # plotting
  
        x = xs.reshape(-1,1)
        r = res.reshape(-1, 1)
        p = pred.reshape(-1, 1)
          
        x = range(len(x))
          
        plt.clf()
        plt.plot(x, r, 'r', x, p, 'b--')
        plt.ylim((-1.2, 1.2))
        plt.draw()
        plt.pause(0.3)  # 每 0.3 s 刷新一次
          
        # 打印 cost 結果
        if i % 20 == 0:
            print('cost: ', round(cost, 4))

　　得到結果：

 

二、CNN主要知識點

　　待整理。