RNN处理长句子

处理文本、获取到X和Y

import tensorflow as tf
import numpy as np
from tensorflow.contrib import rnn
from tensorflow.contrib import layers, seq2seq

tf.set_random_seed(777)

sentence = ("if you want to build a ship, don't drum up people together to "
            "collect wood and don't assign them tasks and work, but rather "
            "teach them to long for the endless immentsity of the sea.")

char_set = list(set(sentence)) # 得到不重复的字母
char_dict = {w : i for i, w in enumerate(char_set)} # 获得词典，以便以字母得到对应的数值

hidden_size = len(char_set) # 隐藏层个数
num_classes = len(char_set) # 分类
sequence_length = 20 # 每次训练的字符串长度为10，数量任意
learning_rate = 0.01 # 学习率


'''
每次循环20个字母
'''
dataX = []
dataY = []
for i in range(0, len(sentence) - sequence_length):
    # 获取特征和标签集数据
    x_str = sentence[i : i + sequence_length]
    y_str = sentence[i + 1 : i + sequence_length + 1]

    # 根据字典获取数据对应数值
    x = [char_dict[i] for i in x_str]
    y = [char_dict[i] for i in y_str]

    # 将数值向量放入集合
    dataX.append(x)
    dataY.append(y)

占位符，session会话中传值调用

X = tf.placeholder(tf.int32, [None, sequence_length])
Y = tf.placeholder(tf.int32, [None, sequence_length])

one-hot编码

X_one_hot = tf.one_hot(dataX, num_classes)

RNN使用

# LSTM单元格，使用hidden_size(每个单位输出向量大小)创建一个lstm单元格
def lstm_cell():
    cell = rnn.BasicLSTMCell(hidden_size, state_is_tuple=True)
    return cell

# RNN单元格
multi_cells = rnn.MultiRNNCell([lstm_cell() for _ in range(2)], state_is_tuple=True)

outputs, _ = tf.nn.dynamic_rnn(multi_cells, X_one_hot, dtype=tf.float32)

添加一层隐藏层，增加深度，为了更准确

X_for_fc = tf.reshape(outputs, [-1, hidden_size])
outputs = layers.fully_connected(X_for_fc, num_classes, activation_fn=None)

# 确定输出的维度
batch_size = len(dataX)
outputs = tf.reshape(outputs, [batch_size, sequence_length, num_classes])

使用了一层全连接之后效果提升很大

下面就是很正常的优化训练

weights = tf.ones([batch_size, sequence_length])

sequence_loss = seq2seq.sequence_loss(logits=outputs, targets=Y, weights=weights)
mean_loss = tf.reduce_mean(sequence_loss)
train_op = tf.train.AdamOptimizer(learning_rate).minimize(mean_loss)

# 打开会话
sess = tf.Session()
sess.run(tf.global_variables_initializer())

# 训练过程
for i in range(500):
    _, loss, results = sess.run(
        [train_op, mean_loss, outputs], feed_dict={X : dataX, Y : dataY}
    )
    for j, result in enumerate(results):
        index = np.argmax(result, axis=1)
        print(i, j, ''.join([char_set[t] for t in index]), loss)

最后输出预测值，输出值中间都为重复值，所以进行判断组合成语句(可以选择从头取，也可以从尾取)。

results = sess.run(outputs, feed_dict={X : dataX})
for j, result in enumerate(results):
    index = np.argmax(result, axis=1)
    if j is 0:
        print(''.join([char_set[t] for t in index]), end='')
    else:
        print(char_set[index[-1]], end='')

预测效果挺好，基本与样本语句符合