embedding_attention_seq2seq

Jump to bottom

Higepon Taro Minowa edited this page Jul 9, 2017 · 3 revisions

Summary

Embed encoder_input
Compute encoder_output state vector using static-rnn and encoder_input
Attach the encoder_output in attention_states, this is used for attention later
Embed decoder_input
Run attention_decoder and compute encoder_outputs using last encoder_states

Code with comments