question about the basic_rnn_seq2seq (the untied architecture)

[Yutasq]

Hi una-dinosauria,

Thank you for your code. I have a little question about your code (line 135-136 in seq2seq_model.py), based on my understanding, it seems that the same cell parameters are used for the encoder and decoder in the untied mode? Should we copy the cell e.g. copy.deepcopy(cell) as written in the basic_rnn_seq2seq that posted by tensorflow, and employ it in the rnn_decoder?

Thanks.