Please read this as it is very important. PLEASE GENERATE WITH THIS! IT'S THE BEST OCHACO MODEL FROM MY TESTING!
There are a lot, and I mean a lot of Ochaco loras on civitai. But I've tested every single one and gotten inconsistent and undesirable results. I really like Ochaco, so I couldn't stand for this any longer. I created my own using a massive data set and 10,000 steps of training. Out of the images i've provided in the listing I never had to tweak or delete or redo any. I encourage you to give it a try. I want to thank every person who has made ochaco loras, and I by no means will ask you to stop using theirs. But from using every Ochaco Lora on the site, nothing has been better in my opinion. I encourage you to generate and share your images so I can see where specifically the model struggles and find a fix for it.
Note for experienced lora trainers: I have the hardware to train locally, but not the workflow or the know how. Please send me a chat if you can help me.
Here are the triggers so you can customize as you'd like:
Mandatory: uwuraka, blush stickers
Hero Costume: herosuit
Hero Costume if you desire her helmet equipped: helmet
UA gym attire: gymuniform
UA school outfit: schooluniform
Cheerleader Outfit: cheerleader
Cheerleader was not present in many frames so expect inconsistencies
Because the dataset is so huge, you can make her wear pretty much whatever you want like grim reaper robes and scythe, have fun and be sure to share your creations!
***PLEASE NOTE** this model performs with NSFW but I did not intend for this to be an nsfw model, so you will need to use the Negative prompt "gaping vagina" or you will get weird results.
If you want results like mine it's always best to use tags such as anime screencap, masterpiece, best quality, extremely detailed
CFG 7 Euler a
Please generate so I can get data to improve the model.
UPDATE I have received reports that civitais on-site generator is not allowing users to select illustrious xl 2 stable for generations. As someone who generates locally I don't understand why this is happening, but the model was trained on Illustrious XL v2 stable you must use that model to achieve consistent and flawless results.