Hi. Thank you for sharing great works.
I have some question related to the memory issues.
After training GON for some dataset (e.g. MNIST), latent vector (z) should be saved for the inference for reconstructing the whole dataset?
I think latent vector with the size of (number of image x nz x 1 x 1) is not small.
Is it right ?
Hi. Thank you for sharing great works.
I have some question related to the memory issues.
After training GON for some dataset (e.g. MNIST), latent vector (z) should be saved for the inference for reconstructing the whole dataset?
I think latent vector with the size of (number of image x nz x 1 x 1) is not small.
Is it right ?