I would prefer if you guys don't use this model for Commercial Use.You can check out my pixiv where I use my lora to create images.<a target="_blank" rel="ugc" href="https://www.pixiv.net/en/users/34413372">https://www.pixiv.net/en/users/34413372</a>v3 (pony sdxl)I used the same images from v2, but updated the txt file.The recommended prompt is adam_blade_76, sun glasses, sunglasses, belt, open clothes, white hair, jacked, coat, open coat, collar, bindi, abs, muscular, muscle, long jacked, holy crose (yes, I wrote it wrong).v2in this version I used 95 images and they are screenshots that I took from the anime called needless. The file is bigger and more flexible. Value between 0.5-0.7 works best. you need to up scaling and down scaling with inpaing to fix the bad faces. to fix the faces, it is recommended to have upscale (highres.fix) between 1.5 (denoising 0.65 max)-2 (denoising 0.55 max)recommended to use these keywords: sun glasses, long jacked, muscle, white hair spiky hair, choker, black glasses,v2 betaunfinished, but you can try it out. struggle with making good faces. try to upscale normally (no img2img or highres.fix). after that inpaint the part of body you don't like and downscale it with img2img inpainting.v1I used around 15 images and trained the lora with anylora model. I recommend anything 4.5 model.value around 0.6-0.8 give good results on that lora.use prompt like these to get the character: black jacket, glasses, belt, white hair, muscle, a man,

adam blade (needless)

adam blade76-000080.safetensors

Files with SHA256: 5fbe3b64453618b60f81a61dac259db24a694ff62cee90f136e2d7ef89efef09