Make concept LoRAs that really teach generalized concepts. <span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">[2026-11-11]<span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">Trained dildo under clothes on MOSS Image, this is my first try, and it still need some work to make it easier to use. [2026-11-4] <span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">Reworked dataset (now with over 300 unique images in anime/real life/3d rendering, detail captioned) and training methods, fixed a issue that caused V2.0 only works well on base model. V3.0 is now checked against many other fine-tuned models and turbo models. <span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">Support for more models incoming. Most cases use this LoRA at a weight of 1.0, which should be fine.<s>For use with Turbo LoRA/checkpoints, please try a higher weight if the default value is not sufficient (e.g., 1.5).</s> Now with Turbo support.Anima-Turbo 1.1 should also be fine with a weight of 1.0. [2026-9-6] Your best dildo under clothes/ dildo under panties LoRA is back!Retrained with a 2-stage training strategy: larger tagged dataset as stage1 (50% steps) and refined tag-natural language mixed dataset as stage2 (50% steps). The LoRA supports more details (clothing, fabric deformation) now!Now also supports anal dildo insertion !<pre><code>New usage (dildo position): 
dildo under clothes, anal object insertion
dildo under clothes, vaginal object insertion</code></pre> [2026-5-19] Added anima-base-1.0 version [2026-2-7] Updated, added an anima-preview version, take it as a preview lora for a preview model. But having natural language processing ability has so much fun! Use your creativity, anima can do so much more than illustrious. Check showcase images to see it's amazing abilities!<span style="background-color:rgb(26, 27, 30);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">For anima, pony style quality tags are needed. (score_9,score_8,score_7,year 2024, masterpiece) <span style="background-color:rgb(26, 27, 30);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">Anima-preview is so good that even though it hasn't finished training yet, it's already one of the best open-source anime models out there. [2024-11-10] Added Illustrious XL version. Completely reworked dataset, now this lora can do more than just a "bulge".Under this line is instruction for Pony version.This model will add a bulge to your character's underwear. Notice: if dildo popped out, try [dildo] under clothes (add '[]' to word dildo ). Trained with 216 images. Tried to train this lora on sd1.5 before, but the results were nothing close to this one. I think it's time to switch to sdxl now.Made by <a target="_blank" rel="ugc" href="https://civarchive.com/user/OnlyWinningMove">OnlyWinningMove</a>

[concept]dildo under clothes xl

Much harder to train than Anima, still improving

toybox_moss_v5_000003500.safetensors

Moss Image Ver 0.1

Anima-Base V3.0

Files with SHA256: cf6ea32dd11b647e61c9224047604a85fe72647a5ebacfe061c23a4f4c6261bb