<a target="_blank" rel="ugc" href="https://github.com/KohakuBlueleaf/LyCORIS">Various version of LoRA </a>of Matikanefukukitaru (umamusume). Put LoRA file in your stable-diffusion-webui/models/lora folder and write LoRA notation in prompt to applying LoRA. &lt;lora:your_loha_name:weight&gt;Thank you Shiraoki-Sama!All of my works is free. If you'd like to support me, feel free to <a target="_blank" rel="ugc" href="https://www.buymeacoffee.com/mhtLoRA">buy me a coffee</a>. 👍Recommend options<ul><li>LoRA weight around 0.8</li><li>Trigger words<ul><li>base<ul><li>matikanefukukitaru \(umamusume\)</li></ul></li><li>Racing competition costume :<ul><li>school uniform, serafuku, blue sailor collar, clothing cutout, shoulder cutout, short sleeves, white shirt, red neckerchief, backpack, bead bracelet, maneki-neko, blue skirt, pleated skirt, ema, ribbon trim, ribbon-trimmed legwear, zettai ryouiki, white thighhighs, shoes, brown footwear, loafers</li><li>(optional) crystal ball, daruma doll, +_+</li></ul></li><li>Full armor kukukitaru :<ul><li>official alternate costume, yellow headwear, beret, jewelry, necklace, bag, bag charm, backpack, puffy long sleeves, bead bracelet, collared dress, yellow dress, frills, charm \(object\), white socks, green shoes</li></ul></li><li>SSR Speed clothes:<ul><li>official alternate costume, hair flower, choker, frilled sleeves, wide sleeves, long sleeves, frills, japanese clothes, green kimono, floral print, hakama, green kimono, sash, obi, shide, yellow skirt, plaid skirt, green socks</li><li>(optional) ema, omikuji</li></ul></li><li>Casual clothes :<ul><li>collared shirt, pink shirt, long sleeves, open clothes, green jacket, brown pants, brown shorts</li><li>or brown skirt, plaid skirt</li></ul></li><li>Tracen school uniform :<ul><li>tracen school uniform, purple shirt, pleated skirt, puffy short sleeves, white skirt, puffy sleeves, summer uniform, frilled skirt, sailor collar, sailor shirt, miniskirt, frills, white thighhighs</li></ul></li></ul></li></ul>Use a model derived from or mixed with animefull model.<ul><li>Settings<ul><li><a target="_blank" rel="ugc" href="https://github.com/AUTOMATIC1111/stable-diffusion-webui">Stable Diffusion WebUI by Automatic1111</a></li><li>DPM++ Series(SDE Karras, 2M Karras, etc.)</li><li>about 20 steps, CFG scale 3.5~6.5</li><li>I recommend <a target="_blank" rel="ugc" href="https://huggingface.co/AIARTCHAN/aichan_blend/blob/main/vae/BerrysMix.vae.safetensors">kl-f8-anime2.vae</a></li><li>CLIP skip = 1</li><li>Use Hires.fix to get higher quality image<ul><li>Up-scaler : Latent series</li><li>Denoising strength 0.50~0.65</li></ul></li></ul></li></ul>LoRA training info<ul><li>trained on <a target="_blank" rel="ugc" href="https://github.com/kohya-ss/sd-scripts">sd-scripts by kohya_ss</a>.<ul><li><a target="_blank" rel="ugc" href="https://github.com/KohakuBlueleaf/LyCORIS">LyCORIS by KohakuBlueleaf</a></li><li>Thank you all a lot!</li></ul></li><li>based on <a target="_blank" rel="ugc" href="https://huggingface.co/a1079602570/animefull-final-pruned/blob/main/model-001.ckpt">Animefull-pruned model</a></li><li>144 images of matikanefukukitaru<ul><li>4x low quality 48 images</li><li>8x medium quality 43 images</li><li>16x high quality 53 images</li><li>=&gt; 1 epoch = 1384 images</li></ul></li><li>replaced character feature tags with matikanefukukitaru \(umamusume\)<ul><li>horse girl, horse tail, brown hair, yellow eyes, etc.</li></ul></li><li>resolution : 768x768 with aspect ratio bucketing.</li><li>other training settings is included in training_data.zip</li></ul>This LoRA may be compatible with animefull-based models (ex. AbyssOrangeMix Series...)This LoRA can generate nsfw images of matikanefukukitaru.All uploaded images are generated by <a target="_blank" rel="ugc" href="https://huggingface.co/WarriorMama777/OrangeMixs/blob/main/Models/AbyssOrangeMix3/AOM3A1.safetensors">AOM3A1</a>Settings of all preview images are same, only version of LoRA differs.And, special thanks to my best friend. DH.

Matikanefukukitaru (umamusume)

LoRA version.<ul><li>dynamically determined rank version<ul><li>original rank is 16. </li><li>both res block and transformer block trained.</li><li>dynamic resizing method is sv_fro, param is 0.985</li><li>It saved 55% of original file size. (37.2MB -&gt; 20.2MB)</li></ul></li><li>Actually, I don't know what is better version :P</li></ul>

Files with SHA256: aa25b8e380a59ba0495f68d30c9855501c98457d76b698b92f9c823d6216bf2e