# Exploring Transformer Backbones for Image Diffusion Models

Source: [Exploring Transformer Backbones for Image Diffusion Models](<https://www.sota2.com/research/paper/68fa6e4b1e209131cea3f15d>)

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

Data updated: 2025-10-26T00:54:03.791Z

Authors: Princy Chahal

Published: 2022-12-27T07:05:14.000Z

## Abstract

We present an end-to-end Transformer based Latent Diffusion model for image synthesis. On the ImageNet class conditioned generation task we show that a Transformer based Latent Diffusion model achieves a 14.1FID which is comparable to the 13.1FID score of a UNet based architecture. In addition to showing the application of Transformer models for Diffusion based image synthesis this simplification in architecture allows easy fusion and modeling of text and image data. The multi-head attention mechanism of Transformers enables simplified interaction between the image and text features which removes the requirement for crossattention mechanism in UNet based Diffusion models.

## 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/2212.14678>)

[PDF](<https://arxiv.org/pdf/2212.14678.pdf>)