Tbh the only reason im doing this is so I can skip training the school uniforms into future characters (........when I get around to them)<s>It....... sort of kind of sort of not really sort of kind of sort of works as a style LoRA? completely unintentional though so don't expect good results. This LoRA is focused on the outfits not the style. THIS IS NOT A STYLE LORA</s>The <a target="_blank" rel="ugc" href="https://civarchive.com/models/22530">trainer I use</a> trains at Clip Skip 2, but the LoRA seems to work well enough with Clip Skip 1, so I uploaded images using both.Trained on "animefull-final-pruned-fp16.safetensors", Example images generated on Anything-V3.<pre><code>WinterBlazer, [red necktie], [red skirt], [breast pocket], [black blazer]</code></pre><pre><code>SummerShortSleeve, [red necktie], [red skirt], [[school uniform]], [tucked In], [tucked In shirt], [black stripes], [white shirt]

- breast pocket</code></pre><pre><code>[gym uniform], [[red trims]], [red gym shorts], [white shirt], [v-neck]</code></pre>While the previews were generated at a strength of 0.9 I think 0.8 is slightly better. Hindsight moment.

Files with SHA256: a67045eaa25dbe7c27ebc3fed4a37bb5ec4c89bf25f9697888ea01321aab175c