2 ms·
I think variational autoencoders (which explicitly model the compression step and the decompression step) seem like the more classical way to do it.
by gcr 11y ago
I think variational autoencoders (which explicitly model the compression step and the decompression step) seem like the more classical way to do it.