This is a new LORA model, which is better than the previous generation of Aerith model dataset. It completely abandoned the previous dataset, so it did not continue to be updated under the original Aerith model. In other words, this model deliberately avoids Overfit during training, weakens character attributes, makes Aerith more realistic and can also bring more creativity.
the Chilout Mix for the Asian face can still effectively output the results.
Careful friends may find that this Aerith is more like Aerith in CG.
So,create as much as you can. This Aerith is very plastic.
I am very looking forward to seeing different Aerith created by you!