Hi,

I've begun to use flux to generate images (using an online version without training any loras) and even though the results are league above the bing image creator I still can't manage to get the hang of covering the face, hair or the clothes in the pics.

Some pics for reference, I really like the gameshow setup has you can see.

Flux's underlying model just isn't trained with enough messy images particularly if you want to cover the face and hair. There isn't anything you can do about this other than train a model. The model I'm testing is doing reasonably well and will be training a follow up soon.

I don't know if this will help in flux or not, but using bing I described the mess and the coverage in the prompts, and less about the face/hair/outfit of the model. Most of my pics I would only use 'woman' to describe the model, and had better luck getting some good coverage. Of course the results varied. best of luck~

Unfortunately the generation service you used hasn't embedded the prompt so I can't build upon your prompt.

But the model will have a lot of biases to avoid obscuring the face as that is "undesirable" by mainstream people. Messg also hits the nail on the head that the model has very little knowledge of WAM.

Try adding something like "her face is obscured by a thick sticky coating of white foam", it seems somewhat sensitive to your word choice as I've had issues with "her body is covered in thick sticky syrup" vs "her body is covered with thick sticky syrup". I haven't tried it yet but I suspect it may even be reactive to "is" vs "was". Another place where I have had some success is "sticky syrup is cascading down over her head and over her face"