Type | |
Stats | 181 22 |
Reviews | (40) |
Published | Jan 13, 2024 |
Base Model | |
Training | Steps: 1 Epochs: 5 |
Usage Tips | Clip Skip: 1 Strength: 1 |
Trigger Words | odamifuu |
Hash | AutoV2 D33F358DFA |
This a model inteded to recreate the art style of Oda Mifuu, an impasto-style with a moe twist.
Use short prompts, with general colors or themes. Objects may confuse and make regular anime art instead of impasto painting.
The Lora is a little finicky and hard to use. Results can vary wildly. With anime models you may want to use weights of 2 or higher even to get close to original results, but I have had little success with this. Euler a produces more consistently good, low-detail results while DPM++2M Karras creates more detailed results with lower quality.
This is the only Lora I will make that I will suggest you use pruned-emaonly or another general diffusion model. Anime models want to make it too detailed. Experiment for cool results. Please show me your results with this Lora!