This repository contains the project course of Natural Language Understanding (NLU) (University of Trento) for the year 2022.
TASK
The task is the creation of a language model using:
- RNN architectures and/or
- transformers
DATASET
BASELINE
- Vanilla: 140 PP
- LSTM: 90.7 PP
A complete report can be found here.
Tensorboard logs can be parsed using this script, it uses a python utility that can be found here.