fcame can be trained as unsupervised Predict new categories?

[watertianyi]

Can fcame be trained as unsupervised to predict new categories? Because fcmae has a coding and decoding structure, it first masks the area and then predicts the area. It is somewhat similar to the characteristics of GAN generation. Can it predict unknown categories?

the weights I trained myself mismatch for stem.0.weight: copying a param with shape torch.Size([64, 4, 4, 4]) from checkpoint, the shape in current model is torch.Size ([64, 3, 4, 4]).

Original downsample_layers.0.0.weight torch.Size([64, 3, 4, 4])
Mine is downsample_layers.0.0.weight torch.Size([64, 4, 4, 4])