Richard Liao

Thanks for the explanation, Ziwei! Can you further elaborate a bit how to convert from a vector (1 dimension) to x and y coordinates which are 2 dimension? Further, I...

It is so long ago that I don't remember it anymore. But you can see in my comment that you should be able to run the commands from the original...

Keras layer interface probably has changed. You need to adapt to the new protocol. Check Keras topology.py file.

I have been using IMDB data released by Kaggle - https://www.kaggle.com/c/word2vec-nlp-tutorial/data For Yelp dataset, you can go to google group discussion: https://groups.google.com/forum/#!topic/keras-users/IWK9opMFavQ. Ben did provide a link for the download.

every sequence of LSTM output is a 2D, and the context vector is 1D. The product of them is 1D. The context vector is trained to assign weight to the...

Which equation are you referring to? The tanh activation at my code refers to equation (5) and (8). h_it is from GRU output.

Ha, you found a HUGE bug in my code that I didn't realize. I'm quite sure you are the first one to point out even someone asked why I use...

Hi aaronrsun, very nice to hear this little repo helped you with your project. If you like, you can cite my Github and this project. And It will be great...

Sorry, I have to be clear that I am not the author of the paper. I was merely interested in the paper and implemented it since the original paper didn't...

I used Keras TimeDistributed(Dense()) layer to implement this verticle neurons interaction of LSTM output before feeding to attention layer. However, I really doubt if it's useful.