Disclaimer

If you generate imagery with my content, and you'd like to share it, please consider using the "Add Post" button at the bottom of the gallery instead of Post Images button at the top of the site, it gives the creator buzz and it costs nothing on your end. The article that was linked is a very important one I recommend reading.

I used to generate images with people's stuff and then use Post Images in bulk and let the resources be auto-detected. I assumed that they are getting buzz when stuff shows up on their page.. nope, they don't. You have to manually post it to the model page's gallery.

I mainly create for fun, but I would appreciate it regardless.

Prompt Usage

Here is what I think you should know about each LoRA. I didn't include quality tags, expressions, poses, and the other stuff. Reference sample images for those.

Important Notes

• Recommended LoRA Strength is about the 0.8 range.

• Trying to remove headwear? Use missing headwear. A lot of outfits with headwear that cover the head include images with it taken off. Could be useful.

• If you're going for anime styles like my samples, put 3d, realistic, game screenshot in negatives.

• If you use the hair styles, colors or eyes in the section above, you might generate my player character.

Training Parameters

All outfits have 50 images containing specific poses, angles and framing, using the exact same dataset. They are trained on very similar settings across both versions, with Illustrious having 15 epochs for 1500 steps and Anima having 10 epochs for 1000 steps.

If a model still has a trigger word with english words or its Illustrious version is still at 1000 steps, it hasn't been re-trained or need to be.. yet.

Load my safetensor into LoRA Metadata Viewer if you're curious about any settings, honestly. The metadata should be there, except for the rare few that used AI Toolkit.