深度学习TextLSTM的tensorflow1.14实现示例
我是王大你是谁 人气:0对单词最后一个字母的预测
LSTM 的原理自己找,这里只给出简单的示例代码,就是对单词最后一个字母的预测。
# LSTM 的原理自己找,这里只给出简单的示例代码 import tensorflow as tf import numpy as np tf.reset_default_graph() # 预测最后一个字母 words = ['make','need','coal','word','love','hate','live','home','hash','star'] # 字典集 chars = [c for c in 'abcdefghijklmnopqrstuvwxyz'] # 生成字符索引字典 word2idx = {v:k for k,v in enumerate(chars)} idx2word = {k:v for k,v in enumerate(chars)} V = len(chars) # 字典大小 step = 3 # 时间步长大小 hidden = 50 # 隐藏层大小 dim = 32 # 词向量维度 def make_batch(words): input_batch, target_batch = [], [] for word in words: input = [word2idx[c] for c in word[:-1]] # 除最后一个字符的所有字符当作输入 target = word2idx[word[-1]] # 最后一个字符当作标签 input_batch.append(input) target_batch.append(np.eye(V)[target]) # 这里将标签转换为 one-hot ,后面计算 softmax_cross_entropy_with_logits_v2 的时候会用到 return input_batch, target_batch # 初始化词向量 embedding = tf.get_variable("embedding", shape=[V, dim], initializer=tf.random_normal_initializer) X = tf.placeholder(tf.int32, [None, step]) # 将输入进行词嵌入转换 XX = tf.nn.embedding_lookup(embedding, X) Y = tf.placeholder(tf.int32, [None, V]) # 定义 LSTM cell cell = tf.nn.rnn_cell.BasicLSTMCell(hidden) # 隐层计算结果 outputs, states = tf.nn.dynamic_rnn(cell, XX, dtype=tf.float32) # output: [batch_size, step, hidden] states: (c=[batch_size, hidden], h=[batch_size, hidden]) # 隐层连接分类器的权重和偏置参数 W = tf.Variable(tf.random_normal([hidden, V])) b = tf.Variable(tf.random_normal([V])) # 这里只用到了最后输出的 c 向量 states[0] (也可以用所有时间点的输出特征向量) feature = tf.matmul(states[0], W) + b # [batch_size, n_class] # 计算损失并进行迭代优化 cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits_v2(logits=feature, labels=Y)) optimizer = tf.train.AdamOptimizer(0.001).minimize(cost) # 预测 prediction = tf.argmax(feature, 1) # 初始化 tf init = tf.global_variables_initializer() sess = tf.Session() sess.run(init) # 生产输入和标签 input_batch, target_batch = make_batch(words) # 训练模型 for epoch in range(1000): _, loss = sess.run([optimizer, cost], feed_dict={X:input_batch, Y:target_batch}) if (epoch+1)%100 == 0: print('epoch: ', '%04d'%(epoch+1), 'cost=', '%04f'%(loss)) # 预测结果 predict = sess.run([prediction], feed_dict={X:input_batch}) print([words[i][:-1]+' '+idx2word[c] for i,c in enumerate(predict[0])])
结果打印
epoch: 0100 cost= 0.003784
epoch: 0200 cost= 0.001891
epoch: 0300 cost= 0.001122
epoch: 0400 cost= 0.000739
epoch: 0500 cost= 0.000522
epoch: 0600 cost= 0.000388
epoch: 0700 cost= 0.000300
epoch: 0800 cost= 0.000238
epoch: 0900 cost= 0.000193
epoch: 1000 cost= 0.000160
['mak e', 'nee d', 'coa l', 'wor d', 'lov e', 'hat e', 'liv e', 'hom e', 'has h', 'sta r']
加载全部内容