The aim of this project is to perform POS tagging as sequence labelling using Recurrent Neural Architectures (RNN)
The focus is on the f1-score of four different models
- Bidirectional LSTM (Bi-LSTM)
- Bidirectional GRU (Bi-GRU)
- Two Bi-LSTM in sequence
- Bi-LSTM and a Dense layer
The last layer of all models is a dense layer used as classifier