SOTAVerified

Very Deep VAEs Generalize Autoregressive Models and Can Outperform Them on Images

2020-11-20ICLR 2021Code Available1· sign in to hype

Rewon Child

arXivPDF

Code Available — Be the first to reproduce this paper.

Reproduce

Code

Abstract

We present a hierarchical VAE that, for the first time, generates samples quickly while outperforming the PixelCNN in log-likelihood on all natural image benchmarks. We begin by observing that, in theory, VAEs can actually represent autoregressive models, as well as faster, better models if they exist, when made sufficiently deep. Despite this, autoregressive models have historically outperformed VAEs in log-likelihood. We test if insufficient depth explains why by scaling a VAE to greater stochastic depth than previously explored and evaluating it CIFAR-10, ImageNet, and FFHQ. In comparison to the PixelCNN, these very deep VAEs achieve higher likelihoods, use fewer parameters, generate samples thousands of times faster, and are more easily applied to high-resolution images. Qualitative studies suggest this is because the VAE learns efficient hierarchical visual representations. We release our source code and models at https://github.com/openai/vdvae.

Tasks

Image Generation

Benchmark Results

DatasetModelMetricClaimedVerifiedStatus
FFHQ 1024 x 1024Very Deep VAEbits/dimension2.42—Unverified
FFHQ 256 x 256Very Deep VAEbits/dimension0.61—Unverified
ImageNet 32x32Very Deep VAEbpd3.8—Unverified
ImageNet 64x64Very Deep VAEBits per dim3.52—Unverified

Reproductions