Stable Diffusion XL

Stability AI · July 2023

activeOpen Weightdiffusionimage
Parameters6.6B
VariantsBase, Refiner, Turbo

Description

Major upgrade to Stable Diffusion with 6.6 billion parameters and a two-stage generation process — a base model creates the image, then a refiner model enhances fine details. Uses two text understanding systems working together to better interpret complex prompts, producing much higher quality 1024×1024 pixel images.

Key Innovations

Diffusion
DiffusionGenerates outputs by gradually denoising random noise into coherent images/audio. The backbone of Stable Diffusion and DALL·E.
Text-to-Image
Text-to-ImageGenerating images from text descriptions — the technology behind DALL·E, Midjourney, and Stable Diffusion.
Open Weight
Open WeightModel weights are publicly released but training data/code may not be. Enables fine-tuning but not full reproduction.

Family Tree

Lineage

Stable Diffusion 1.5Stable Diffusion XL

Successors (1)

Related Research (1)

2020 · UC Berkeley

Showed that gradually adding noise to data and then learning to reverse the process could generate images rivaling GANs, with more stable training and…