BalancedRandomSampler

[srguo24]

TATS/tats/data.py

Line 281 in 03bbce1

if hasattr(self.args, 'balanced_sampler') and self.args.balanced_sampler and train:

In the above line there's a BalancedRandomSampler class that does not seem to be defined or imported anywhere, which causes an error when I try to create the dataloader with the corresponding args.

I see in the sampling script that for class conditioning, the model should generate an equal number of samples for each class when calculating FVD. Should this class balance also be enforced for the true samples?