A bug in the code may be the reason

[agcala]

Continuing the discussion from Multi-class classification:

@Rojin

test_loader_X = torch.utils.data.DataLoader(dataset=tr_latent_X,
                                             batch_size=bs,
                                             shuffle=False)

It seems to me there is a bug in the code this should have said

test_loader_X = torch.utils.data.DataLoader(dataset=te_latent_X,
                                             batch_size=bs,
                                             shuffle=False)

[ptrblck]

That might be the case. Would you mind continuing the discussion in the linked topic and ping the author, in case she is not aware of this typo?