welcome to our DL final project :)

To run the model, run python model.py
See the following optional arguments:
--load_weights: takes a string representing the name of the model (in model_checkpts) that you want to load
--save_weights: takes a string representing the name you want to give the model, which will be the name of the folder it is saved in in model_checkpts
--test: will test the model instead of training; takes a string representing the name of the model (in model_checkpts) that you want to test
--test_gui: will allow the user to provide an input, runs the model on that input and prints the output
--bert: indicates the model should use BERT instead of distilBERT (note this will use a lot of time and memory)
--from_titles: indicates you want to use the data generated from paper titles rather than rephrased from human-written abstracts
--batch_sze: takes an int of the batch size you wish to use; defaults to 256 --num_epochs: takes an int of the number of epochs you want to train the model for; defaults to 10
--percent_data: takes a float of the proportion of data you wish to use for training and testing; defaults to 1
--max_num_tokens: takes an int of the max number of tokens to use when tokenizing inputs; defaults to 512