Stable Diffusion XL Refiner 1.0
Specialized SDXL checkpoint for the final denoising and image refinement stage
Stable Diffusion XL Refiner 1.0
Specialized SDXL checkpoint for the final denoising and image refinement stage
Stable Diffusion XL Refiner 1.0 Overview
Stable Diffusion XL Refiner 1.0 is the refinement checkpoint in the SDXL 1.0 pipeline. It is trained for the final denoising steps on existing SDXL latents or image-to-image inputs, making it useful when a base SDXL generation needs cleaner fine detail, stronger local texture, and a more polished final pass rather than a full standalone generation workflow.
More models from Stability AI
Stable Diffusion XL v1.0 VAE Fix is the SDXL 1.0 base checkpoint packaged with a corrected VAE for more stable inference. It keeps SDXL's strong prompt understanding, broad visual range, native 1024x1024 generation, and support for both text-to-image and image-to-image workflows, while reducing the artifact and decoding issues associated with the original embedded VAE.