Kim Seonghyeon

I have tried to train on FFHQ.

No, model in the paper is too large to use in my environments. In my cases I got 45% training accuracies for top level codes.

@wwlCape If you want to try 1024 model then you need to use bottom + middle + top models, and larger pixelsnail model. But I don't know this repository can...

@Chauncy-Cai PixelSNAIL is quite heavy model. Actually, the default model setting in this respository is light version of the model specified in the paper.

It is in the https://github.com/rosinality/vq-vae-2-pytorch/tree/master/distributed. I don't know why torch.distributed is used, instead of this.

Could you check free disk spaces? If it is enough then you can try to reduce map_size.

It will be depend on jpeg compression rates, but I don't think it will affect the quality much.

Latent code is compressed representation of images.

You don't need save it. Size of latent codes is 32x32 & 64x64 with 512 discrete codes, so size of latent codes will be 9 * (32 * 32 +...

Of course it will be reduced if you uses additional compression.