ITADN

FLUX 1 Fill Pro

#508Openkindratyshyn-bot 创建于 2026-02-15
Experimenting with this model and it tends to insert unwanted objects when I just want it to naturally extend existing context. For example like the two unsightly buildings structures on the sides, or a strange cave exit when it should simply extend the landscape (attached). Varying guidance parameters and prompts doesn't seem to help. Please advise or suggest a different outpainting model. I am giving inputs with diagonal corners or irregular parts of the image missing, which, depending on the image, the model seems to be handling well. <img width="1080" height="1440" alt="Image" src="https://github.com/user-attachments/assets/77def00c-48b2-46a7-88e3-a35e70a975ce" /> <img width="1439" height="1063" alt="Image" src="https://github.com/user-attachments/assets/f460d541-aa4f-420f-9862-1b24a83dcc22" />
0 条评论