boogu-project/Boogu-Image

Waiting for Edit Turbo

期待ing...

Might be provided in 1-2 weeks : ) Depends on whether there is any other emergent issue.

Thanks for your interest. We are still cooking our code. It depends on how many people need the training code and finetuning code.

> Thank for your advice. Will do.

Any detail plan?

Thank for your advice. Will do.

我试加入了turbo lora，可实现cfg1.0四步出图，画质降了少许。

> One image for now. Maybe two images after two weeks. Any other advice?

In my experience, three input image edits are more than enough.

会尽快出

Let me add that this model is very special for some specific tasks. It's one of the best I've seen in regards to what I call "character identity bleeding", where most model, including the most recent or even costly paid models will merge or "bleed" the physical attributes of a subject into all characters in the scene. Because of this quality, it's also very good at isolating the position of characters individually and when interacting with each other, for creating more complex scenes. However, the one thing this model needs the most is fine-tunning and lora capabilities! I'm sure the open-source community would make this model shine brightly if given the chance.

so Boogu-Image, its not perform like Flux klein with 4 images, then i need to check upcoming "JoyAI-Image-Edit-Plus" model. anyways thanks for the info.

One image for now. Maybe two images after two weeks. Any other advice?