Don't expect it to be perfect, that's it.
Support me: https://ko-fi.com/elevatedone
Recommendations Pony:
LoRA weight: 0.6-1.0
Negative prompts: Use negative embedding for pony.
If you use "red cape", it will be there, if not, then not. The same for the nail (use "sword" in prompt, although it's not really stable).
Since it was trained on the same dataset as SD 1.5 (r34 mostly) and Pony can do ???? much better
If you want her to be ?????, use "?????, black body" in the positive
If you want her to be not ?????, use "black bodysuit" in positive (or something alike) and "?????, ????" in the negative
Recommendations SD 1.5:
LoRA weight: 0.6-0.8, 0.8 is the most stable
Do not use long prompts, it creates disfigurement
Negative prompts: Deep Negative V1.75T and Bad Hands 5 TIs, prompts like: signature, multiple views are also recommended
Positive: can be anything, just not too much
Description
Trained on around 300 images which were picked by hand from multiple image boards
FAQ
Details
Files
Available On (2 platforms)
Same model published on other platforms. May have additional downloads or version variants.