Just a deep learning library for my convenience to develop the code quickly.
- k-max pooling: https://discuss.pytorch.org/t/resolved-how-to-implement-k-max-pooling-for-cnn-text-classification/931/3 , http://www.cnblogs.com/Joyce-song94/p/7277871.html
- Connect to Stanford core NLP and support automatically annotate corpus for training
- Vis idea, use your own layer to support this feature
- training APP (support results distribution)
- support pair wise training (hinge loss, max margin loss)