NVAE: A Deep Hierarchical Variational Autoencoder · HackerPapers discussions

NVAE: A Deep Hierarchical Variational Autoencoder — mentioned by kelseyfrog

Fri, 27 Jun 2025 03:35:05 GMT

Forgotten from like 2021? NVAE[1] was a great paper but maybe four years is long enough to be forgotten in the AI space? shrug 1. NVAE: A Deep Hierarchical Variational Autoencoder https://arxiv.org/pdf/2007.03898

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

NVAE: A Deep Hierarchical Variational Autoencoder — 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.

NVAE: A Deep Hierarchical Variational Autoencoder — HN submission

Mon, 21 Sep 2020 22:10:54 GMT

NVAE: A Deep Hierarchical Variational Autoencoder

7 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

NVAE: A Deep Hierarchical Variational Autoencoder — mentioned by aaron-santos

Mon, 21 Sep 2020 22:01:23 GMT

Great timing. Just about to get into SE blocks after reading through Nvidia's NVAE paper[1]. [1] NVAE: A Deep Hierarchical Variational Autoencoder https://arxiv.org/abs/2007.03898

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