Any support is appreciated! This will go towards better hardware.
https://buymeacoffee.com/tucker_dev
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 Karras
Steps: 30-60
CFG: 5-8
Suggest adding "comic, anime, ugly, cartoon, watermark" to negative prompt
Description
Feet are much more realistic. Used DPO primarily
Details
Downloads
37
Platform
SeaArt
Platform Status
Available
Created
3/18/2024
Updated
4/1/2024
Deleted
-
Trigger Words:
f33t
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.