suriyadeepan / practical_seq2seq

A simple, minimal wrapper for tensorflow's seq2seq module, for experimenting with datasets rapidly
http://suriyadeepan.github.io/2016-12-31-practical-seq2seq/
GNU General Public License v3.0
570 stars 270 forks source link

Repeating few tokens in all outputs: #78

Open irdanish11 opened 5 years ago

irdanish11 commented 5 years ago

for example, if I ask Q: 'How are you' Answer from Chatbot: ' I gonna gonna gonna gonna gonna look look look look look much much much much' I changed the vocabulary size to 8000 and used 100,000 epochs but still the results were the same.