Regarding training_opt: {open_threshold: 0.1}

[rahulvigneswaran]

What does training_opt: {open_threshold: 0.1} do? It looks like, it is pointing to the theta below but it is not being used at all.

def F_measure(preds, labels, theta=None):
    # Regular f1 score
    return f1_score(labels.detach().cpu().numpy(), preds.detach().cpu().numpy(), average='macro')

[KaihuaTang]

I built my project on top of https://github.com/facebookresearch/classifier-balancing, which itself is built on top of https://github.com/zhmiao/OpenLongTailRecognition-OLTR

It's common to have some deprecated codes in the project. As long as, it doesn't cause any bug

[rahulvigneswaran]

@KaihuaTang Understood. Thanks alot.