使用paddle进行酒店评论的情感分类8——网络训练与评估

对paddle上的案例进行少许调整得到如下，剩下的运行即可

# 库文件导入
# encoding=utf8
import re
import random
import requests
import numpy as np
import paddle
from paddle.nn import Embedding
import paddle.nn.functional as F
from paddle.nn import LSTM, Embedding, Dropout, Linear
import os
import jieba
import paddle.fluid

from build_dict import build_dict
from convert_corpus_to_id import convert_corpus_to_id
from data_preprocess import data_preprocess
from load_comment import load_comment
from __init__ import SentimentClassifier
from build_batch import build_batch

paddle.set_device('gpu')

# 这之后的PaddlePaddle计算将会在GPU上进行

paddle.seed(0)
random.seed(0)
np.random.seed(0)

train_corpus =  load_comment(True)
test_corpus = load_comment(False)
train_corpus = data_preprocess(train_corpus)
test_corpus = data_preprocess(test_corpus)
word2id_freq, word2id_dict = build_dict(train_corpus)
train_corpus = convert_corpus_to_id(train_corpus, word2id_dict)
test_corpus = convert_corpus_to_id(test_corpus, word2id_dict)

# 定义训练参数
epoch_num = 5
batch_size = 128

learning_rate = 0.0001
dropout_rate = 0.2
num_layers = 2  # 这里改为2
hidden_size = 256
embedding_size = 256
max_seq_len = 128
vocab_size = len(word2id_freq)

# 实例化模型
sentiment_classifier = SentimentClassifier(hidden_size, vocab_size, embedding_size,  num_steps=max_seq_len, num_layers=num_layers, dropout_rate=dropout_rate)

# 指定优化策略，更新模型参数
optimizer = paddle.optimizer.Adam(learning_rate=learning_rate, beta1=0.9, beta2=0.999, parameters= sentiment_classifier.parameters()) 

# 定义训练函数
# 记录训练过程中的损失变化情况，可用于后续画图查看训练情况
losses = []
steps = []

def train(model):


    # 开启模型训练模式
    model.train()
    
    # 建立训练数据生成器，每次迭代生成一个batch，每个batch包含训练文本和文本对应的情感标签
    train_loader = build_batch(word2id_dict, train_corpus, batch_size, epoch_num, max_seq_len)
    
    for step, (sentences, labels) in enumerate(train_loader):
        # 获取数据，并将张量转换为Tensor类型
        sentences = paddle.to_tensor(sentences)
        labels = paddle.to_tensor(labels)
        
        # 前向计算，将数据feed进模型，并得到预测的情感标签和损失
        logits = model(sentences)

        # 计算损失
        loss = F.cross_entropy(input=logits, label=labels, soft_label=False)
        loss = paddle.mean(loss)

        # 后向传播
        loss.backward()
        # 更新参数
        optimizer.step()
        # 清除梯度
        optimizer.clear_grad()

        if step % 100 == 0:
            # 记录当前步骤的loss变化情况
            losses.append(loss.numpy()[0])
            steps.append(step)
            # 打印当前loss数值
            print("step %d, loss %.3f" % (step, loss.numpy()[0]))

#训练模型
train(sentiment_classifier)

# 保存模型，包含两部分：模型参数和优化器参数
model_name = "sentiment_classifier"
# 保存训练好的模型参数
paddle.save(sentiment_classifier.state_dict(), "{}.pdparams".format(model_name))
# 保存优化器参数，方便后续模型继续训练
paddle.save(optimizer.state_dict(), "{}.pdopt".format(model_name))


def evaluate(model):
    # 开启模型测试模式，在该模式下，网络不会进行梯度更新
    model.eval()

    # 定义以上几个统计指标
    tp, tn, fp, fn = 0, 0, 0, 0

    # 构造测试数据生成器
    test_loader = build_batch(word2id_dict, test_corpus, batch_size, 1, max_seq_len)
    
    for sentences, labels in test_loader:
        # 将张量转换为Tensor类型
        sentences = paddle.to_tensor(sentences)
        labels = paddle.to_tensor(labels)
        
        # 获取模型对当前batch的输出结果
        logits = model(sentences)
        
        # 使用softmax进行归一化
        probs = F.softmax(logits)

        # 把输出结果转换为numpy array数组，比较预测结果和对应label之间的关系，并更新tp，tn，fp和fn
        probs = probs.numpy()
        for i in range(len(probs)):
            # 当样本是的真实标签是正例
            if labels[i][0] == 1:
                # 模型预测是正例
                if probs[i][1] > probs[i][0]:
                    tp += 1
                # 模型预测是负例
                else:
                    fn += 1
            # 当样本的真实标签是负例
            else:
                # 模型预测是正例
                if probs[i][1] > probs[i][0]:
                    fp += 1
                # 模型预测是负例
                else:
                    tn += 1

    # 整体准确率
    accuracy = (tp + tn) / (tp + tn + fp + fn)
    
    # 输出最终评估的模型效果
    print("TP: {}\nFP: {}\nTN: {}\nFN: {}\n".format(tp, fp, tn, fn))
    print("Accuracy: %.4f" % accuracy)

# 加载训练好的模型进行预测，重新实例化一个模型，然后将训练好的模型参数加载到新模型里面
saved_state = paddle.load("./sentiment_classifier.pdparams")
sentiment_classifier = SentimentClassifier(hidden_size, vocab_size, embedding_size,  num_steps=max_seq_len, num_layers=num_layers, dropout_rate=dropout_rate)
sentiment_classifier.load_dict(saved_state)

# 评估模型
evaluate(sentiment_classifier)