A RNN based Part of Speech Tagger for hieroglyphic unicode. The models have pretty low accuracy, 60-70%, but this isn't bad considering that it is a character level RNN.
Implementing other types of taggers, using them together, trying to improve performance.
Code by Jordan Furutani Uses PyTorch, training and testing data from Simon Schweitzers AES [https://github.com/simondschweitzer/aes].