This Lora is supposed to produce funny cartoon faces.
I've trained it with screencaps from various cartoon TV shows.
In V1, apparently there were a few too much screencaps with Pete from Disney's Goofy Troop, as that seems to be all that V1 produce.
V2 seems more versatile (but I'm still not quite happy with it and will probably make another version sometime).
For the time being I recommend playing with the weights of this Lora and whatever character Lora you may be using. You should still be able to get some good results.