How to fix faces of trained loras.
I have been training a few loras lately and have been wondering what is the best way to fix faces? If you look at my models i have uploaded the eyes and mouths are a little wonky.
its the result of your dataset, you can't do it unless you use other Lora to rewrite the texts for that particular prompt. can also be because of Overfitting. you have to train it again according to the results rn.