I'm back with a brand new LoRa. I wanted to test myself with the Ahegao, and I've achieved impressive results with this model, but a few errors can occur with this version. Like the fact that the model tends to generate a lot of tattoos on the bodies, this can be reduced or even removed by adding "tatoo, tattoo" to your negative prompt. I'll try to correct this problem in a future version. In the meantime I haven't really fallen victim to any other negative elements with regard to this model.
