Theano Implementation of DMN+ (Improved Dynamic Memory Networks) from the paper by Xiong, Merity, & Socher at MetaMind, http://arxiv.org/abs/1603.01417 (Dynamic Memory Networks for Visual and Textual Question Answering)
System expects the maximum length of input/questions in train and test data to be the same. Otherwise it raises a dimension error. I figured out this while I was trying with my own data.
Good to consider both training and test set while deciding the max question/input length.
System expects the maximum length of input/questions in train and test data to be the same. Otherwise it raises a dimension error. I figured out this while I was trying with my own data.
Good to consider both training and test set while deciding the max question/input length.