Illustrious info : Innate style of the model feels not enough trained, lora should reinforce it. Probably a little overbaked though. Config with params attached for this version.Pony info : Two versions, one trained from base pony and have some nice aesthetics, while suffering at style sometimes, second is trained from autismmix and forcing a style replication really hard, but almost corrupts small details and anatomy, either use pony version with higher than 1 weight if style doesn't replicated enough or just a autismmix version for a harder style replication. Personally I like more the pony version with rare need to raise the weight. <a target="_blank" rel="ugc" href="https://github.com/Haoming02/sd-webui-vectorscope-cc">Vectorscope </a>is also a good addition

Files with SHA256: 560b1ea24987595b0fa4880b967afe816b1c93c1c030a4b7b1dece95da0c6395