Load pre-trained embedding weights. Arguments: vectors {torch.Tensor} -- from "TEXT.vocab.vectors".

Arguments: vector_size {int} -- word embedding size. vocab_size {int} -- the size of the vocabulary. hidd

读取分类目录，固定

读取词汇表

生成批次数据

Compute the number of predictions that match some target classes in the output of a model. Args: output_probabilities: A tensor o

常用文件操作，可在python2和python3间切换. mode: 'r' or 'w' for read or write

将文件转换为id表示

迭代器，将数据分批传给模型

Apply a masked softmax on the last dimension of a tensor. The input tensor and mask should be of size (batch, *, sequence_length). Args:

pad_sequences 把序列长度转变为一样长的，如果设置了maxlen则长度统一为maxlen，如果没有设置则默认取 最大的长度。填充和截取包括两种方法，post与pre，post指从尾部开始处理，pre指从头部 开始处理，默认都是从尾部开始。 Argumen

读取文件数据

Compute the loss and accuracy of a model on some validation dataset. Args: model: A torch module for which the loss and accuracy must

Apply a weighted sum on the vectors along the last dimension of 'tensor', and mask the vectors in the result with 'mask'. Args: t

测试集上准曲率评估

测试集上准确率评估

生成藏头诗

生成藏头诗

模式一：随机生成诗歌 模式二：生成藏头诗 模式三：给定首句生成诗 :return:

自由生成一首诗歌

提取tfidf特征

获取已经使用的时间

读入json文件，返回诗句list，每一个元素为一首诗歌(str类型表示)

常用文件操作，可在python2和python3间切换. mode: 'r' or 'w' for read or write

code from keras Pads each sequence to the same length (length of the longest sequence). If maxlen is provided, any sequence longer th

获取原数据并预处理 :param data_path: 数据存放的路径 :param category: 数据的类型 :param author: 作者名称 :param constrain: 长度限制 :return: list ['床前明

分词去停用词

数据预处理

预测结果

提供数据

读取类别

读取数据

将文本的类别写到本地

对文本进行处理，取出脏数据

Train a model for one epoch on some input data with a given optimizer and criterion. Args: model: A torch module that must be tra

rnn_type must be a class inheriting from torch.nn.RNNBase

Arguments: vector_size {int} -- word embedding size. vocab_size {int} -- vocabulary size. Keyword Arguments:

Arguments: input_size {int} -- input size to the feedforward neural network. output_size {int} -- output size of the

Arguments: input_size {int} -- input size to the feedforward neural network. output_size {int} -- output size of the

根据字典kwargs 更新 config参数

根据训练集构建词汇表，存储

Apply dropout to the input batch of sequences. Args: sequences_batch: A batch of sequences of vectors that will serve

Args: premise_batch: A batch of sequences of vectors representing the premises in some NLI task. The batch is ass

Apply dropout to input tensor. Parameters ---------- input_tensor: ``torch.FloatTensor`` A tensor of shap

Arguments: x {torch.Tensor} -- input tensor with shape [batch_size, seq_length]

Arguments: x {torch.Tensor} -- input embeddings with shape [batch, seq_len, input_size] Returns: output {torc