KallMeKris

This was my third attempt at making a LoRA, I just never published it. It was trained using Linaqruf's LoRA google collab.

I seem to get the best results using dreamlike, ICBITP, hardblend, juggernaut, and cyberrealistic. DPM++SDE, unipc, and Euler a, and 20-30 steps

All feedback is welcome and would love to see your creations using this!

It is trained on and intended for realistic models, it is not anime friendly.

This lora works best at a weight of 0.75-1.0.