<h2 id="single-outfit-support">Single-Outfit Support</h2>This model is trained only on Mint's default outfit.Single costume training captured complex clothing details better than multi-costume training.If you're looking for the multi-outfit version, check the below 👇<a target="_blank" rel="ugc" href="https://civarchive.com/models/2996298/mintnte3outfit">https://civarchive.com/models/2996298/mintnte3outfits</a><hr /><h3 id="quick-reference">Quick Reference</h3>LoRA Weight: <code>0.8</code> – <code>1.0</code>Character Trigger:<pre><code>mint \(nte\), cat girl, cat ears, animal ear fluff, long hair, multicolored hair,streaked hair, aqua hair, braid, cat tail, pink eyes, large breasts</code></pre>Outfit1: Default<pre><code>black gloves, hairband, (cat face on hairband:0.8), white over shirt, black dress, black corset, black skirt, white thighhighs,thigh strap, black boots, fanny belt pouch, gold aiguillette</code></pre>Managed to land on an art style right between 3D and the official art.Feel free to add "3D" to the NP if you prefer.<hr /><h3 id="my-reserch">My reserch</h3>Outfit1 Key Points A and B<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/79d8f014-607e-4ebb-a00c-62e0ba44eac1/original=true/79d8f014-607e-4ebb-a00c-62e0ba44eac1.jpeg" />single outfit models outperformed 3outfits models.As a side effect, having too much 3D source data caused the output style to look more like 3D.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/4346bea0-ed92-483d-9cb0-e49ee4cf54d9/original=true/4346bea0-ed92-483d-9cb0-e49ee4cf54d9.jpeg" /><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/75420af5-c6b9-4e75-b1af-40f0b12fd154/original=true/75420af5-c6b9-4e75-b1af-40f0b12fd154.jpeg" /><hr />Comparing Anima and ILThere is no difference between them.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/173fac1b-dd8a-45f8-961b-b52d877b81ee/original=true/173fac1b-dd8a-45f8-961b-b52d877b81ee.jpeg" />Surprisingly, Anima performed just as well as IL even though the text encoder wasn't a perfect match.Turns out, as long as the dataset quality is solid, training just the U-Net is more than enough.

Mint_NTE_Outfit1

12 images datasetDim:8, Alpha:8

Files with SHA256: 831e24702951b416e49892654b62f1836f16d029bf9bba40b1d38e0789264a74