Implements optuna optimization algorithms using sklearn ML algorithms (currently, GradientBoostingClassifier, SVC, Neural Net, and ElasticNet.) Accuracy, auROC, or auPRC of predictions on cross-validation datasets can be used as the optimization objective.
Datasets (i.e. the crossvalidation, training, and testing set(s)) are user-specified according to the python class specified in config_function.py.
python3 run_optuna_training_testing.py --lang_model_type Rostlab_Bert
python3 run_optuna_training_testing.py --model_name SVC --scoring_metric auROC --lang_model_type Rostlab_Bert
To find all possible arguments to the main function run_optuna_training_testing.py, please see run_optuna_args.py.