question about config

[wmpauli]

Hi,

Thank you for sharing your code on github. I have a question regarding the configs.

In the config for ImageNet and CelebA, there are two lines that seem to be more relevant for the lsun datasets.
lsun_categories_train: [bedroom_train]
lsun_categories_test: [bedroom_test]

Are they simply ignored in these cases, because they are not using the lsun dataset?

Also, for celebA and the lsun dataset examples, does nlabels=1 indicate that there is only one sample per category.

Lastly, for lsun_bridges, it looks like training and testing is done on the same data (_train). Is that intentional?

[LMescheder]

Yes, lsun_categories_train and lsun_categories_test are ignored when not training on LSUN. nlabels=1 simply indicates that there are no categories for celebA.
For your last question: we currently only have a train set, as we currently only measure inception score, which does not require a separate test set.