SHI-Labs / Prompt-Free-Diffusion

Prompt-Free Diffusion: Taking "Text" out of Text-to-Image Diffusion Models, arxiv 2023 / CVPR 2024

Home Page:https://arxiv.org/abs/2305.16223

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

One reference image one model?

shang106121 opened this issue · comments

Hi, interesting work!
But i have a question,for different reference images, do we need to train a new model for each reference image?Is each trained SeeCoder only suitable for a specific reference image?

No, SeeCoder is train free. Moreover, it can be adapted to many T2I diffusers.