Using Recurrent Neural Networks to predict text using the Alice in Wonderland dataset from Project Gutenberg
3
stars
13
forks
source link
on 2 layer Bidirectional LSTM model, changed optimizer from "rmsprop" to "adam", increased hidden size and batch size to next power of 2(256), accuracy=89% #12
Bidirectional-2layer-LSTM.txt