comfy workflow please?
comfy workflow please?
I found via search this:
https://github.com/boogu-project/ComfyUI-Boogu/tree/main
But i have not installed yet ... i wait if we get native support in comfyUI
Have a look at https://github.com/Comfy-Org/ComfyUI/pull/14523
I did have to check out that pull request to get the boogu clip type.
The pullrequest has now been merged!
omnigen2 as clip type works too.
beyond that it a bog standard workflow like Z-image and similar.
Many people encounter inability to install the TextEncodeBooguEdit node
this has actually sneakily become my most currently used image model. Certainly of anything released since Flux.2 Klein. The hotfixed versions have fixed a few issues. And it is decent at using reference images for new scene with edit, and quite good at just making them with the main model. Although base fp8+turbo lora seems to be much better than the actual turbo model.
And it just works! It doesn't need a fricking script as prompt, it knows how many arms and legs a person has and where they should be... And the image quality doesn't beat Flux.2 but deffo hidreamO1. Only real letdown is precise image editing tasks.