Semantic-Typing-Of-Events

Setup

conda create --name step-seq2seq python=3.7.10
conda activate step-seq2seq
pip install -r requirements.txt 

Run

python main.py --model-path model_path --type-model type_model --batch-size batch_size --only-test only_test --max-len max_len

model_path is the path where you want to save the model, you can to choose between: models/their_split, models/their_split_mixed, models/their_split_all_processes.

type_model is the model of BART, you can to choose between base or large.

batch_size is size of the batch, default is 2

only_test is the boolean for only testing the model, default is False

max_len used to distinguish between with and without glosses, default is 20 refers to without glosses and 175 is with glosses

for example for training phase:

python main.py --model-path models/their_split_all_processes --type-model large --batch-size 2

for example for only testing phase:

python main.py --model-path models/their_split_all_processes --type-model large --batch-size 2 --only-test True