Loads a vocabulary file into a dictionary.

Args: vec: size=(batch_size, vanishing_dim, hidden_dim) m_size: hidden_dim Returns: size=(batch_size, hidden_dim)

feats: size=(batch_size, seq_len, tag_size) mask: size=(batch_size, seq_len) tags: size=(batch_size, seq_len)

:param path:数据文件路径 :param max_length: 最大长度 :param label_dic: 标签字典 :return:

Do the forward algorithm to compute the partition function (batched). Args: feats: size=(batch_size, seq_len, self.targe

Args: scores: size=(seq_len, batch_size, tag_size, tag_size) mask: size=(batch_size, seq_len) tags: size=

Args: feats: size=(batch_size, seq_len, self.target_size+2) mask: size=(batch_size, seq_len) Returns:

Args: feats: size=(batch_size, seq_len, tag_size) mask: size=(batch_size, seq_len) tags: size=(batch_size

random initialize hidden variable

默认保留所有模型 :param model: 模型 :param path: 保存路径 :param loss: 校验损失 :param last_loss: 最佳epoch损失 :param kwargs: every_epoch or best_

Args: target_size: int, target size use_cuda: bool, 是否使用gpu, default is True average_batch: bool, loss是否作

args: sentence (word_seq_len, batch_size) : word-level representation of sentence hidden: initial hidden state