Usually i wouldn't post anything i trained because it's kinda tricky to handle and extremely rigged when it comes to prompts or which model works. Also a lot of plugins for 1111 play a big role sometimes. They work for me but maybe not for others.

Just a little test to see what i need to change in order to make easier to use and more compatible and flexible. I know that the .txt files and source pictures need a bit more curating, but like i said, it works for me.

This one here was made for a friend ( and for myself ofc ) that asked me about it. There is a model on Civitai that pretty much has the same contend, but it's SD1.5. Old meme is old, but i like the random results you get. Some LoRAs I've trained pretty much make no sense, but they change the composition and style of an image drastically at lower value... like most LoRAs i guess.

Don't mind some of the words i've used in the example images. Most of the time they did nothing, but it steered the image into a certain direction. Just for my own entertainment and to see what happens.

I will probably use this model for further training.

-Trained on BaseXL

-BF16