I tried implementing the seq to seq model for chatbot with Cornell movie dialogue dataset. I used a single lang class for both the input and output. Except that everything is the same. Even after 10000 epochs, the error is not going under 3.5. Any suggestions?
I tried implementing the seq to seq model for chatbot with Cornell movie dialogue dataset. I used a single lang class for both the input and output. Except that everything is the same. Even after 10000 epochs, the error is not going under 3.5. Any suggestions?