Classification with imbalance dataset

[leminhnhat007]

Hello,
I am trying to apply the classification of the KAN model to the imbalanced dataset. The initial idea was to define two functions: train_f1_score and test_f1_score to calculate the f1 score for the train and test output. Then apply metrics(train_f1_score, test_f1_score) instead of metrics=(train_acc, test_acc) for the KAN model. However, it KAN.py looks like the metrics are only used to append into the results but do not have any role in training the model:

    if metrics != None:
        for i in range(len(metrics)):
            results[metrics[i].__name__] = []

Is there any other way for KAN to deal with an imbalanced dataset?
Thank you