# High Fidelity Image Synthesis With Deep VAEs In Latent Space

Source: [High Fidelity Image Synthesis With Deep VAEs In Latent Space](<https://www.sota2.com/research/paper/68fa6ee61e209131cea42bc1>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-26T01:08:40.604Z

Authors: Troy Luhman, Eric Luhman

Published: 2023-03-23T23:45:19.000Z

## Abstract

We present fast, realistic image generation on high-resolution, multimodal datasets using hierarchical variational autoencoders (VAEs) trained on a deterministic autoencoder's latent space. In this two-stage setup, the autoencoder compresses the image into its semantic features, which are then modeled with a deep VAE. With this method, the VAE avoids modeling the fine-grained details that constitute the majority of the image's code length, allowing it to focus on learning its structural components. We demonstrate the effectiveness of our two-stage approach, achieving a FID of 9.34 on the ImageNet-256 dataset which is comparable to BigGAN. We make our implementation available online.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2303.13714>)

[PDF](<https://arxiv.org/pdf/2303.13714.pdf>)