selected model image

Free Service

Astria

Train and personalize a stable diffusion model with your own dataset

Finetune Stable Diffusion on objects or people with Astria. Their algorithm is based on Nataniel Ruiz's paper DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generation↗︎. The Dataset for finetuning can be as small as 10-20 pictures of the desired object or person.