Open e-yi opened 3 years ago
https://github.com/vandit15/Class-balanced-loss-pytorch/blob/921ccb8725b1eb0903b2c22a1a752a594fcae138/class_balanced_loss.py#L28
should be alpha: A float tensor of size [num_classes]
alpha: A float tensor of size [num_classes]
@e-yi refer to https://github.com/fcakyon/balanced-loss for improved and maintained version
https://github.com/vandit15/Class-balanced-loss-pytorch/blob/921ccb8725b1eb0903b2c22a1a752a594fcae138/class_balanced_loss.py#L28
should be
alpha: A float tensor of size [num_classes]