def forward()

in models.py [0:0]

• 26 lines of code
• 7 McCabe index (conditional complexity)

    def forward(self, sent_tuple):
        # sent_len: [max_len, ..., min_len] (bsize)
        # sent: (seqlen x bsize x worddim)
        sent, sent_len = sent_tuple

        # Sort by length (keep idx)
        sent_len_sorted, idx_sort = np.sort(sent_len)[::-1], np.argsort(-sent_len)
        sent_len_sorted = sent_len_sorted.copy()
        idx_unsort = np.argsort(idx_sort)

        idx_sort = torch.from_numpy(idx_sort).cuda() if self.is_cuda() \
            else torch.from_numpy(idx_sort)
        sent = sent.index_select(1, idx_sort)

        # Handling padding in Recurrent Networks
        sent_packed = nn.utils.rnn.pack_padded_sequence(sent, sent_len_sorted)
        sent_output = self.enc_lstm(sent_packed)[0]  # seqlen x batch x 2*nhid
        sent_output = nn.utils.rnn.pad_packed_sequence(sent_output)[0]

        # Un-sort by length
        idx_unsort = torch.from_numpy(idx_unsort).cuda() if self.is_cuda() \
            else torch.from_numpy(idx_unsort)
        sent_output = sent_output.index_select(1, idx_unsort)

        # Pooling
        if self.pool_type == "mean":
            sent_len = torch.FloatTensor(sent_len.copy()).unsqueeze(1).cuda()
            emb = torch.sum(sent_output, 0).squeeze(0)
            emb = emb / sent_len.expand_as(emb)
        elif self.pool_type == "max":
            if not self.max_pad:
                sent_output[sent_output == 0] = -1e9
            emb = torch.max(sent_output, 0)[0]
            if emb.ndimension() == 3:
                emb = emb.squeeze(0)
                assert emb.ndimension() == 2

        return emb