Training on "black sclera".
I have been working to integrate various unique features into LoRa's (and now LoCon). I have found making individual trainings on each feature, then merging the ones needed together gets remarkable results.
Eyes are hard to get right, this helps but isn't a perfect solution yet though.
This has been trained on about 80 AI generated images from various models (picked out from hundreds), after using a couple not so great lora's I had experimented with before.
This can be used stand-alone to add black sclera with much higher efficiency, or mixed with other LoRA / LoCon to mix the feature in. I recommend to avoid using this with multiple other LoRA / LoCon without mixing first, it is more prone to cause issues.
If you like this, get me a coffee :D
Description
Should be usable with any anime-based model.
It likes to focus on the face, and use darker clothing. But can get pulled away from it with different prompts/weights.
Still working on the process. So any feedback is appreciated.
FAQ
Comments (3)
its pretty hard to creat a black sclera picture like you did...maybe because my gpu is too old?
Try with the same model I used. GPU should not be much of a factor unless it's an AMD one.
Exactly, like felor said - try starting with the model used for the previews. I tried a variety of them - and yeah, it "feels" super hard - but once you start dial back the lora strength (or just generate your picture - then use inpainting) -> you'll get somewhat proper results =)
It's definitely useful, especially since without it - most models don't get what you want to achieve (because it lacks training/reference data).