These versions were trained on Moss Image Base. Example images were generated with Moss Image SFT.
These loras were trained on images from Megami Magazine from late 2018 to 2026. Moss already knows the Megami Magazine tag but it’s riddled with japanese text. These loras were trained with the text manually removed and the data curated to present a specific style or look.
To get the results you see in the images above, these are the settings I used.
Visit My Site For Models I Can't Post On CivitAI: https://qtgurl.com/
The loras for Moss Image were trained at 1792 resolution. So they're capable of generating at around 1792x1792 or higher.(I use 1600x1984)
MeMaMoss - V1
Trained on 68 new and old images. Dataset was re-collected and edited for the best possible quality.
Words about Moss
Moss came out of no where and gave us not only danbooru based anime images but it's super uncensored. Just check the example images above to see for yourself. I decided to train a lora on it just to see if it can really bring out the MeMa style and I wasn't disappointed. Not only can this model do good anime but it can do amazing backgrounds, logos, multiple characters and I'm sure much more that I haven't got around to testing. But it's not perfect. It doesn't know all of danbooru, it doesn't know many characters and artist. Hopefully all this will be solved if/when they release a proper anime finetune. Lastly, this is a slow model, I use slow settings and high resolution for my gens, it takes me twice as long to get a picture compared to Anima. But it's worth it in my opinion. The images look so amazing.
I hope you enjoy this lora as much as I am. 2d style for life!