Family Guy Core BETA 1

My first Family Guy model for Anima. It is good when it is able to grab the correct character. Inpainting helps very much. The images presented are not edited or touched up in any way, but it took several attempts and playing with the prompts to get the character correctly.

This LORA does Lois very well and the other characters probably at a C or D level.
It is definitely not the best mix. I think it had too many characters and too few pictures. I will either try to make one with a larger dataset or more likely break the Lords into individual characters or character duos, which seemed to work rather well.

Much of this is experimentation with Anima training so use with caution. It tends to work pretty well with Lois and other characters can definitely be wrestled in


On the other hand, this may be simply a feature of Anima, as prompting is a little bit different structurally. I will either retrain this model or, alternatively, it's been a user error. I'd like to see what other people can do with it, as it seems that if you prompt it well, it does excellently