NLP之文本分类模型调优（模型基于tensorflow1.14）

项目链接：https://pan.baidu.com/s/1yOu0DogWkL8WOJksJmeiPw?pwd=4bsg 提取码: 4bsg 复制这段内容后打开百度网盘手机App，操作更方便哦
--来自百度网盘超级会员v4的分享

1.基于上一篇文章中的文本分类项目进行精度调优，提升模型准确率，需要改动的地方除了参数以外，就是模型结构，这里主要就是进行模型结构调优，因此代码只需要修改rnn_model.py，修改前后代码如下：

原模型结构：

双层LSTM+两层全连接神经网络+softmax

原模型代码：

#!/usr/bin/python
# -*- coding: utf-8 -*-

import tensorflow as tf

class TRNNConfig(object):
    """RNN配置参数"""

    # 模型参数
    embedding_dim = 64      # 词向量维度
    seq_length = 600        # 序列长度
    num_classes = 10        # 类别数
    vocab_size = 5000       # 词汇表达小

    num_layers= 2           # 隐藏层层数
    hidden_dim = 128        # 隐藏层神经元
    rnn = 'gru'             # lstm 或 gru

    dropout_keep_prob = 0.8 # dropout保留比例
    learning_rate = 1e-3    # 学习率

    batch_size = 128         # 每批训练大小
    num_epochs = 10          # 总迭代轮次

    print_per_batch = 100    # 每多少轮输出一次结果
    save_per_batch = 10      # 每多少轮存入tensorboard


class TextRNN(object):
    """文本分类，RNN模型"""
    def __init__(self, config):
        self.config = config

        # 三个待输入的数据
        self.input_x = tf.placeholder(tf.int32, [None, self.config.seq_length], name='input_x')
        self.input_y = tf.placeholder(tf.float32, [None, self.config.num_classes], name='input_y')
        self.keep_prob = tf.placeholder(tf.float32, name='keep_prob')#定义dropout占位符，输入失活比例

        self.rnn()

    def rnn(self):
        """rnn模型"""

        def lstm_cell():   # lstm核
            return tf.contrib.rnn.BasicLSTMCell(self.config.hidden_dim, state_is_tuple=True)

        def gru_cell():  # gru核
            return tf.contrib.rnn.GRUCell(self.config.hidden_dim)

        def dropout(): # 为每一个rnn核后面加一个dropout层
            if (self.config.rnn == 'lstm'):
                cell = lstm_cell()
            else:
                cell = gru_cell()
            return tf.contrib.rnn.DropoutWrapper(cell, output_keep_prob=self.keep_prob)

        # 词向量映射
        with tf.device('/cpu'):
            embedding = tf.get_variable('embedding', [self.config.vocab_size, self.config.embedding_dim])  # [5000, 64]
            embedding_inputs = tf.nn.embedding_lookup(embedding, self.input_x)  # [?, 600, 64]

        with tf.name_scope("lstm"):
            '''
            TF 低版本RNN多层获取cell，使用一下代码
            cell = rnn.BasicLSTMCell(hidden_size, state_is_tuple=True)
            cell = rnn.MultiRNNCell([cell] * 2, state_is_tuple=True)
            '''
            # 多层rnn网络
            cells = [dropout() for _ in range(self.config.num_layers)]
            rnn_cell = tf.contrib.rnn.MultiRNNCell(cells, state_is_tuple=True)

            _outputs, _ = tf.nn.dynamic_rnn(cell=rnn_cell, inputs=embedding_inputs, dtype=tf.float32)  # [?, 600, 128]
            last = _outputs[:, -1, :]  # 取最后一个时序输出作为结果  # [?, 128]

        with tf.name_scope("score"):
            # 全连接层，后面接dropout以及relu激活
            fc = tf.layers.dense(last, self.config.hidden_dim, name='fc1')
            fc = tf.contrib.layers.dropout(fc, self.keep_prob)
            fc = tf.nn.relu(fc)

            fc=tf.layers.dense(fc,units=64,name='fc2')
            fc = tf.contrib.layers.dropout(fc, self.keep_prob)
            fc = tf.nn.relu(fc)

            fc = tf.layers.dense(fc, units=32, name='fc3')
            fc = tf.contrib.layers.dropout(fc, self.keep_prob)
            fc = tf.nn.relu(fc)


            # 分类器
            self.logits = tf.layers.dense(fc, self.config.num_classes, name='fc4')
            self.y_pred_cls = tf.argmax(tf.nn.softmax(self.logits), 1)  # 预测类别

        with tf.name_scope("optimize"):
            # 损失函数，交叉熵
            cross_entropy = tf.nn.softmax_cross_entropy_with_logits(logits=self.logits, labels=self.input_y)
            self.loss = tf.reduce_mean(cross_entropy)
            # 优化器
            self.optim = tf.train.AdamOptimizer(learning_rate=self.config.learning_rate).minimize(self.loss)

        with tf.name_scope("accuracy"):
            # 准确率
            correct_pred = tf.equal(tf.argmax(self.input_y, 1), self.y_pred_cls)
            self.acc = tf.reduce_mean(tf.cast(correct_pred, tf.float32))

修改为双层双向LSTM+4层全连接神经网络+softmax

#!/usr/bin/python
# -*- coding: utf-8 -*-

import tensorflow as tf

class TRNNConfig(object):
    """RNN配置参数"""

    # 模型参数
    embedding_dim = 64      # 词向量维度
    seq_length = 600        # 序列长度
    num_classes = 10        # 类别数
    vocab_size = 5000       # 词汇表达小

    num_layers= 2           # 隐藏层层数
    hidden_dim = 128        # 隐藏层神经元
    rnn = 'gru'             # lstm 或 gru

    dropout_keep_prob = 0.8 # dropout保留比例
    learning_rate = 1e-3    # 学习率

    batch_size = 128         # 每批训练大小
    num_epochs = 10          # 总迭代轮次

    print_per_batch = 100    # 每多少轮输出一次结果
    save_per_batch = 10      # 每多少轮存入tensorboard


class TextRNN(object):
    """文本分类，RNN模型"""
    def __init__(self, config):
        self.config = config

        # 三个待输入的数据
        self.input_x = tf.placeholder(tf.int32, [None, self.config.seq_length], name='input_x')
        self.input_y = tf.placeholder(tf.float32, [None, self.config.num_classes], name='input_y')
        self.keep_prob = tf.placeholder(tf.float32, name='keep_prob')#定义dropout占位符，输入失活比例

        self.rnn()

    def rnn(self):
        """rnn模型"""

        def lstm_cell():   # lstm核
            return tf.contrib.rnn.BasicLSTMCell(self.config.hidden_dim, state_is_tuple=True)

        def gru_cell():  # gru核
            return tf.contrib.rnn.GRUCell(self.config.hidden_dim)

        def bilstm_cell():
            return tf.contrib.rnn.BasicLSTMCell(self.config.hidden_dim)

        def dropout(): # 为每一个rnn核后面加一个dropout层
            if (self.config.rnn == 'lstm'):
                cell = lstm_cell()

            else:
                cell = gru_cell()
            return tf.contrib.rnn.DropoutWrapper(cell, output_keep_prob=self.keep_prob)

        # 词向量映射
        with tf.device('/cpu'):
            embedding = tf.get_variable('embedding', [self.config.vocab_size, self.config.embedding_dim])  # [5000, 64]
            embedding_inputs = tf.nn.embedding_lookup(embedding, self.input_x)  # [?, 600, 64]

        with tf.name_scope("lstm"):
            '''
            TF 低版本RNN多层获取cell，使用一下代码
            cell = rnn.BasicLSTMCell(hidden_size, state_is_tuple=True)
            cell = rnn.MultiRNNCell([cell] * 2, state_is_tuple=True)
            '''
            cell_fw=[dropout() for _ in range(self.config.num_layers)]
            cell_bw=[dropout() for _ in range(self.config.num_layers)]
            lstm_forward = tf.contrib.rnn.MultiRNNCell(cells=cell_bw, state_is_tuple=True)
            lstm_backward=tf.contrib.rnn.MultiRNNCell(cells=cell_fw, state_is_tuple=True)
            outputs, states = tf.nn.bidirectional_dynamic_rnn(cell_fw=lstm_forward, cell_bw=lstm_backward, inputs=embedding_inputs,
                                                              dtype=tf.float32)
            outputs_fw = outputs[0]
            outputs_bw = outputs[1]
            last = outputs_fw[:, -1, :] + outputs_bw[:, 0, :]  # 取最后一个时序输出作为结果  # [?, 128]

        with tf.name_scope("score"):
            # 全连接层，后面接dropout以及relu激活
            fc = tf.layers.dense(last, self.config.hidden_dim, name='fc1')
            fc = tf.contrib.layers.dropout(fc, self.keep_prob)
            fc = tf.nn.relu(fc)

            fc=tf.layers.dense(fc,units=64,name='fc2')
            fc = tf.contrib.layers.dropout(fc, self.keep_prob)
            fc = tf.nn.relu(fc)

            fc = tf.layers.dense(fc, units=32, name='fc3')
            fc = tf.contrib.layers.dropout(fc, self.keep_prob)
            fc = tf.nn.relu(fc)


            # 分类器
            self.logits = tf.layers.dense(fc, self.config.num_classes, name='fc4')
            self.y_pred_cls = tf.argmax(tf.nn.softmax(self.logits), 1)  # 预测类别

        with tf.name_scope("optimize"):
            # 损失函数，交叉熵
            cross_entropy = tf.nn.softmax_cross_entropy_with_logits(logits=self.logits, labels=self.input_y)
            self.loss = tf.reduce_mean(cross_entropy)
            # 优化器
            self.optim = tf.train.AdamOptimizer(learning_rate=self.config.learning_rate).minimize(self.loss)

        with tf.name_scope("accuracy"):
            # 准确率
            correct_pred = tf.equal(tf.argmax(self.input_y, 1), self.y_pred_cls)
            self.acc = tf.reduce_mean(tf.cast(correct_pred, tf.float32))

修改结束后，记得在原项目中的TextRNN\checkpoints\textrnn中将训练好的模型删除，避免冲突。

注：如使用tensorflow gpu版本进行模型训练，将with tf.device('/cpu'):改为with tf.device('/gpu:0'):

注：可通过此命令下载scikit-learn包：conda install -c anaconda scikit-learn 注:(安装sklearn)

注：conda环境:
conda create -n 环境名 python=版本号 : 新建虚拟环境
activate 环境名进入环境
conda deactivate 退出
conda env list 显示环境
conda remove -n 环境名 --all