I am a beginner and recently starting learning and implementing a simple seq2seq
My seq2seq colab
when predicting i am getting the same response I will appreciate your help
Hi @blanky, An LSTM-based encoder-decoder (Seq2Seq) model though properly configured , may generate identical outputs for different inputs if training is insufficient. To resolve this increase the training duration to at least 100 epochs . I have successfully replicated the code without any issue after increasing the epochs from 10 to 100.
I am attaching gist file for your reference. Thank you!
Thanks @Divya_Sree_Kayyuri well my initial accuracy was good enough, so I thought the problem was somewhere else, the new results are better than before thanks for the help.