使用Pytorch进行文本分类——BILSTM+ATTENTION

网络结构

代码实现

class TextBILSTM(nn.Module):
    
    def __init__(self,
                 config:TRNNConfig,
                 char_size = 5000,
                 pinyin_size = 5000):
        super(TextBILSTM, self).__init__()
        self.num_classes = config.num_classes
        self.learning_rate = config.learning_rate
        self.keep_dropout = config.keep_dropout
        self.char_embedding_size = config.char_embedding_size
        self.pinyin_embedding_size = config.pinyin_embedding_size
        self.l2_reg_lambda = config.l2_reg_lambda
        self.hidden_dims = config.hidden_dims
        self.char_size = char_size
        self.pinyin_size = pinyin_size
        self.rnn_layers = config.rnn_layers

        self.build_model()


    def build_model(self):
        # 初始化字向量
        self.char_embeddings = nn.Embedding(self.char_size, self.char_embedding_size)
        # 字向量参与更新
        self.char_embeddings.weight.requires_grad = True
        # 初始化拼音向量
        self.pinyin_embeddings = nn.Embedding(self.pinyin_size, self.pinyin_embedding_size)
        self.pinyin_embeddings.weight.requires_grad = True
        # attention layer
        self.attention_layer = nn.Sequential(
            nn.Linear(self.hidden_dims, self.hidden_dims),
            nn.ReLU(inplace=True)
        )
        # self.attention_weights = self.attention_weights.view(self.hidden_dims, 1)

        # 双层lstm
        self.lstm_net = nn.LSTM(self.char_embedding_size, self.hidden_dims,
                                num_layers=self.rnn_layers, dropout=self.keep_dropout,
                                bidirectional=True)
        # FC层
        # self.fc_out = nn.Linear(self.hidden_dims, self.num_classes)
        self.fc_out = nn.Sequential(
            nn.Dropout(self.keep_dropout),
            nn.Linear(self.hidden_dims, self.hidden_dims),
            nn.ReLU(inplace=True),
            nn.Dropout(self.keep_dropout),
            nn.Linear(self.hidden_dims, self.num_classes)
        )


    def attention_net_with_w(self, lstm_out):
        '''
        :param lstm_out: [batch_size, time_step, hidden_dims * num_directions(=2)]
        :return:
        '''
        lstm_tmp_out = torch.chunk(lstm_out, 2, -1)
        # h [batch_size, time_step, hidden_dims]
        h = lstm_tmp_out[0] + lstm_tmp_out[1]
        # atten_w [batch_size, time_step, hidden_dims]
        atten_w = self.attention_layer(h)
        # m [batch_size, time_step, hidden_dims]
        m = nn.Tanh()(h)
        # atten_context [batch_size, time_step, time_step]
        atten_context = torch.bmm(m, atten_w.transpose(1, 2))
        # softmax_w [batch_size, time_step, time_step]
        softmax_w = F.softmax(atten_context, dim=-1)
        # context [batch_size, hidden_dims, time_step]
        context = torch.bmm(h.transpose(1,2), softmax_w)
        context_with_attn = h.transpose(1, 2) + context
        # result [batch_size, hidden_dims]
        # result = torch.sum(context, dim=-1)
        result = torch.sum(context_with_attn, dim=-1)
        return result


    def forward(self, char_id, pinyin_id):
        # char_id = torch.from_numpy(np.array(input[0])).long()
        # pinyin_id = torch.from_numpy(np.array(input[1])).long()

        sen_char_input = self.char_embeddings(char_id)
        sen_pinyin_input = self.pinyin_embeddings(pinyin_id)

        sen_input = torch.cat((sen_char_input, sen_pinyin_input), dim=1)
        # input : [len_seq, batch_size, embedding_dim]
        sen_input = sen_input.permute(1, 0, 2)
        output, (final_hidden_state, final_cell_state) = self.lstm_net(sen_input)
        # output : [batch_size, len_seq, n_hidden]
        output = output.permute(1, 0, 2)
        # final_hidden_state = torch.mean(final_hidden_state, dim=0, keepdim=True)
        # atten_out = self.attention_net(output, final_hidden_state)
        atten_out = self.attention_net_with_w(output)
        return self.fc_out(atten_out)

Attention计算

1. 将BILSTM网络输出的结果（shape：[batch_size, time_step, hidden_dims * num_directions(=2)]）拆成两个大小为[batch_size, time_step, hidden_dims]的Tensor；
2. 将第一步拆出的两个Tensor进行相加运算得到h（shape：[batch_size, time_step, hidden_dims]）；
3. 将h作为输入，通过self.attention_layer得到attention的计算向量atten_w（shape：[batch_size, time_step, hidden_dims]）；
4. 将第二步的h进行tanh()激活，得到m（shape：[batch_size, time_step, hidden_dims]），留待后续进行残差计算；
5. 将atten_w的2、3维度进行调换，并与m进行矩阵的乘法运算，得到atten_context（shape：[batch_size, time_step, time_step]）;
6. 将atten_context最后一维进行softmax计算，得到当前时刻相对于所有时刻的权重：softmax_w（shape：[batch_size, time_step, time_step]）；
7. 将h的2、3维度进行调换，并与softmax_w进行矩阵运算，得到基于权重的context（shape：[batch_size, hidden_dims, time_step]）；
8. 将h的2、3维度进行调换，并与context进行求和运算，得到context_with_attn（shape：[batch_size, hidden_dims, time_step]）；
9. 将context_with_attn的最后一维进行求和，得到result（shape：[batch_size, hidden_dims]）；
   10.计算结束，返回result；

模型效果

• 1层BILSTM在训练集准确率：99.8%，测试集准确率：96.5%；
• 2层BILSTM在训练集准确率：99.9%，测试集准确率：97.3%；

调参

• dropout的值要在 0.1 以下（经验之谈，笔者在实践中发现，dropout取0.1时比dropout取0.3时在测试集准确率能提高0.5%）。

相关文章

• 使用TextCNN模型进行文本分类（链接）；
• 使用Transformer模型进行文本分类（链接）