static2.py 文件源码

def forward(self, x, lengths):
        batch_size, seq_length = x.size()[:2]

        emb = Variable(torch.from_numpy(
            self.initial_embeddings.take(x.numpy(), 0)),
            volatile=not self.training)
        h = Variable(torch.zeros(batch_size, self.model_dim), volatile=not self.training)

        for t in range(seq_length):
            inp = emb[:,t,:]
            h = self.rnn(inp, h)

        h = F.relu(self.l0(F.dropout(h.squeeze(), 0.5, self.training)))
        h = F.relu(self.l1(F.dropout(h, 0.5, self.training)))
        y = F.log_softmax(h)
        return y