This LoRA is intended to mimic the art style of Plankton Heretic: <a target="_blank" rel="ugc" href="https://bsky.app/profile/did:plc:bpvi7o2nckaog3g3qzzmmvum￼Support">https://bsky.app/profile/did:plc:bpvi7o2nckaog3g3qzzmmvum Support</a> the artist! As they do somewhat niche subjects I am unsure how versatile this LoRA is yet. Trained on 2D images, no line art. Mostly Warcraft characters, and muscular characters, mostly female. PonyXL works like a charm. Ponyrealism will make 3D characters. It occasionally butchers anatomy, so you might have to be specific. I will add more training in the future. Here's what I know so far: A VAE may deep fry it a bit, but I haven't tested all of them. With some work, it will work without the trigger word, but the trigger word helps a lot. Easily does clothing. Penises tend to be uncircumcised if flaccid. It tends to do very simplistic backgrounds, probably because the artist did so few. I am unsure if it occasionally struggles with anatomy, or if that is just AI in general. Draft mode is basically a rorschach test unless I am doing something wrong. Using a lighter weight may give better results. It struggles with the orc teeth, I have no idea why. It may mix colors as well. Does this weird missing penis thing on occasion as well. Blends well with other LoRAs I have tested so far, which can also negate issues it has. It struggles with the orc teeth, I have no idea why. It may mix colors as well. Does this weird missing penis thing on occasion as well. Blends well with other LoRAs I have tested so far. (pony) UPDATE: V1.0B - Added a bit more art to the data set, there really isn't much left I could add that would be beneficial. A few pieces were character sheets so you might get bits from that. (pony) UPDATE: V1.0C - Added a few more art pieces to the data. Also added illustrious V1.0

Files with SHA256: 845858e9a4a1b974fdfef82250eb4efed64701efd80be63e7e14f6cfb03e4aa1