In PyTorch, when loading LSUN images, only a few hardcoded classes are accepted:
https://pytorch.org/vision/stable/datasets.html#lsun
The accepted classes are:
[‘bedroom’, ‘bridge’, ‘church_outdoor’, ‘classroom’,
‘conference_room’, ‘dining_room’, ‘kitchen’,
‘living_room’, ‘restaurant’, ‘tower’]
However, there are also 20 object categories like cars and horses, as shown on the official LSUN website: LSUN
How can we load these object classes instead?