Lora for capturing the Kyabakurabakufu style, strong line art style, can do clothing, but primarily its really good at following Kura's shading techniques on bodies, good for NSFW scenes. Kyabakurabakufu is a NSFW artist so yeah just a heads up. V1 was made on 8dim and V2 was made on 16dim, hence the smaller file size (Its not a virus). Refer to this link if you are still unsure about small LoRA. <a target="_blank" rel="ugc" href="https://civarchive.com/models/6596/luisap-lora-dont-need-to-be-bigger-new-recipe">https://civarchive.com/models/6596/luisap-lora-dont-need-to-be-bigger-new-recipe</a> I used OrangeCocoa5050 to train V1 at 512x768, and V2 at 512x512 For V1 you need to push the weighting a little harder, but for V2 you can sit around .5 - .7 and get a consistent effect if you are not battling with another LoRA for attention. *Do Keep in mind for majority of the Images for V2 I used the Controlnet Extension because its broken how well it makes images. Therefore you can use the prompt data, but you will not get the exact results I have. (Besides maybe the Nami jeans one and the Lisa one) If you have any LoRA requests and wouldnt mind buying me a Hot Chocolate, please consider going to my page. <a target="_blank" rel="ugc" href="https://www.buymeacoffee.com/JohnStinko">https://www.buymeacoffee.com/JohnStinko</a>

Kyabakurabakufu style LORA

16DIM, 288 Images, 57,600 steps, 512x512 (But bucketed this time) It's a much stronger effect

Files with SHA256: f29b1f505d9e5b9f54e4be67b6b0e3d7ee108c545a1a051936395969e2523175