Any support is appreciated! This will go towards better hardware.<a target="_blank" rel="ugc" href="https://buymeacoffee.com/tucker_dev">https://buymeacoffee.com/tucker_dev</a>This is a LoHA model trained off Reddit. Mostly for teaching myself different techniques. Please use responsibly. v1.1 focused on a cleaner dataset.Sampling: DPM++ 2M SDE KarrasSteps: 30-60CFG: 5-8Suggest adding "comic, anime, ugly, cartoon, watermark" to negative prompt

Smokin' Feet

<ul><li>Feet are much more realistic. Used DPO primarily</li></ul>

Files with SHA256: b78787ae6e9fde6951f4cffad6a545f844512d534ccdbe84588e24b698e47d19