/citation_intent_classification

Contrastive learning for intent

Primary LanguagePython

Citation_classification

Data process

We execute

python data_processing

to extract the required parts of the jsonl file into a new csv file.

Search hyperparameters

  1. The run_optuna function is run in the main.py file and is used to find the right combination of hyperparameters.
  2. Record the final combination of hyperparameters output in the function for the next step of training.

Training

Run the main_run function in the main.py file. Train the model on the training set and also generate a file of the prediction results for the test set.

Note

Do not run both the run_optuna and main_run functions in the main.py file.