In decoder mode, the generated summary came out only 1 character?
I followed the instructions to train my models and test, And in decode mode, I set max_dec_step = 1 for running the decoder one step at a time, However the Generated Summary came out only one character. Is there anyone can help me with this issue? I'm not sure if any parameter are not setting or any other problem. Thanks.
In beam_search.py replace while steps < 100 and len(results) < FLAGS.beam_size: instead while steps < FLAGS.max_dec_steps and len(results) < FLAGS.beam_size:
I followed the instructions to train my models and test, And in decode mode, I set max_dec_step = 1 for running the decoder one step at a time, However the Generated Summary came out only one character. Is there anyone can help me with this issue? I'm not sure if any parameter are not setting or any other problem. Thanks.
FLAGS.max_dec_steps=1, means beam_search only one step,
and decode_onestep only return _topk_ids of one character.
so you can try to use larger max_dec_steps