Working with Language Models

This repository uses Pytorch and NLTK libraries to build different neural network models for natural language processing using WikiText-2 database.

RNN

To run the RNN model, the file main_RNN.py has to be executed, which will call other files like vocabulary_RNN and import_data_RNN to run correctly.

LSTM

Similarly to run LSTM model, the file main_LSTM.py has to be executed. LSTM model can be also executed using colab link - https://colab.research.google.com/drive/1_Et2rmyNZ5F-YeKrB7dPHzXPb1SRgNRo

Bonus

To run the RNN, LSTM models incorporating the bonus task, the file bonus_RNN.py and bonus_LSTM.py needs to be executed, and will run again the same models but with some functions implemented manually such as the softmax, logits and negative log-probabilities.