not-so-BigGAN: Generating High-Fidelity Images on Small Compute with Wavelet-based Super-Resolution — mentioned by gwern

Wed, 20 Jan 2021 15:19:54 GMT

If you don't have TFRC access, this is out of the question as a hobbyist. It's >32 GPU-months. Sorry. However, you can still transfer learn this particular model with <1 GPU-month, if you have some anime or artwork dataset you want (eg a specific character). The good news is that there's been huge progress in GAN efficiency over the past few years, and we are rapidly approaching the point where ImageNet-scale datasets can be reasonably trained with 1 GPU-month or less. Worth noting in that vein are VQ-GAN https://compvis.github.io/taming-transformers/ and not-so-BigGAN…

Paper and discussion history. A mention is not an endorsement.

not-so-BigGAN: Generating High-Fidelity Images on Small Compute with Wavelet-based Super-Resolution · HackerPapers discussions

not-so-BigGAN: Generating High-Fidelity Images on Small Compute with Wavelet-based Super-Resolution — mentioned by gwern