tma15 / bunruija Goto Github PK

I'll implement Quasi-Recurrent Neural Network-based classifier.

paper: https://arxiv.org/abs/1611.01576

The current implementation of PRADO is slow mainly because of a string projection.
I'll implement a C++ version of the string projection

Reproduice looks like a typo of reproduce.

To be:

- **Easy to reproduce the training of a model**: because all hyperparameters of a model are stored in a yaml file, it is easy to reproduce the model.

The current implementation of neural network-based classifiers requires large amount of memory.
Thus, I'll implement the following algorithm, which is memory efficient and achieves near state-of-the-art performance on document classification tasks:

PRADO: Projection Attention Networks for Document Classification On-Device

https://www.aclweb.org/anthology/D19-1506/