This is continuation of Taikaponi, which can be found at https://civitai.com/models/428329
This is a Pony-based checkpoint merge made with "Train Difference" towards different checkpoints and Lora merges, instead of regular flat merging... Which messes up things a bit:
some Loras which have been trained with "base pony", have also inherited it's drawing style, so when applied to this model it will mess up the styles, because of the "train difference". Some work, some wont, so it's trial and error and using lora weights.
The example pictures have been made using SD Forge https://github.com/lllyasviel/stable-diffusion-webui-forge
So to start a prompt, use pony basic quality trigger words:
score_9, score_8_up, score_7_up
For negatives, put what you dislike.
Because it's a Pony base, you can also utilize the data selection tags of:
