NN - FBMix

Experimental LORA trained on mixed dataset mostly anime. Most likely need more training but I like the small changes it makes on low strength.

Recommented on 0.5 strength.

Any feedbacks are welcome.