Make LSTM backwards compatible and fix incremental generation

        # compute attention

        attn_scores = (source_hids * x.unsqueeze(0)).sum(dim=2)

        attn_scores = F.softmax(attn_scores.t(), dim=1).t()  # srclen x bsz

        attn_scores = F.softmax(attn_scores.t()).t()  # srclen x bsz

        # sum weighted sources

        x = (attn_scores.unsqueeze(2) * source_hids).sum(dim=0)

        self.fc_out = Linear(out_embed_dim, num_embeddings, dropout=dropout_out)

    def forward(self, input_tokens, encoder_out):

        if self._is_incremental_eval:

            input_tokens = input_tokens[:, -1:]

        return self._forward(input_tokens, encoder_out)

    def _forward(self, input_tokens, encoder_out):

        bsz, seqlen = input_tokens.size()

        # get outputs from encoder