>>102635191
Its quite literally no different than generating the image. You pick a model, add your loras, type up a prompt, add an starting image and run it.
The only thing that can be confusing is that the lora situation which more impact than just adding characters or styles, we use some loras to distill the model to make it run faster/less steps (kinda like the hyper loras for Stable Diffusion) (PUSA, Light2xv) but I promise i'm not doing much more than any image genner.