Recommended weight: 0.6 - 1
Recommended CFG scale: 4.5 - 8
img2img denoising strength: 0.25 - 0.4
Steps: 15 - 30
LoRA was trained on 251 images, tagged mostly by WD ViT Tagger v3 and partially by me. I'll probably make a v2 once I gather more high quality data and if people like v1.
Any feedback or suggestions are welcome :D