pytorch自然语言处理基础模型之一：NNLM

任务描述：pytorch实现简单的神经网络语言模型NNLM

一、数据

sentences=[‘i like dog’, ‘i love coffee’, ‘i hate milk’]

二、目标

当输入[‘i’,‘like’], [‘i’,‘love’], [‘i’,‘hate’]时，分别预测下一个单词为[‘dog’], [‘coffee’]，[‘milk’]

三、完整代码

1. 导入需要的库，设置数据类型

import torch
import torch.nn as nn
import torch.optim as optim
from torch.autograd import Variable
dtype = torch.FloatTensor

2. 分词，创建字典

sentences = ['i like dog', 'i love coffee', 'i hate milk']
words = ' '.join(sentences).split()
words_dict = list(set(words))
num2word = {
   index:word for index, word in enumerate(words_dict)}
word2num = {
   word:index for index, word in enumerate(words_dict)}

3. 准备数据

def make_batch(sentences):
    input_batch = []
    target_batch = []
    for sen in sentences