rlhmns - experimental

rl-hmns.safetensors

v1.0

Shared originally on <a target="_blank" rel="ugc" href="https://publicprompts.art/real-humans-sdxl-lora/">PublicPrompts</a>strength of "1.0" is good for general use, if the faces are not real enough, increase the weight. Also closeups with extremely realistic features, need higher weight than 1 Trigger word is "photo" and "portrait photo" forces closeups specially on higher weight I see huge potential for merging, but didn't test yet, it's more of an experiment that went way better than expected so I shared it Edit: - Just noticed that's it's so good for selfie prompts "selfie portrait photo of..." at strength 1.5 - also noticed that I'm adding images with no prompts, feel free to ask for any missing one, I don't mind sharing

Files with SHA256: 1e55b965bd8b058ec393d9331c479cbf7e8e03e18387c45f2f319ecb3a30e8fb