图片为本人生成,使用的模型和均为C站公开模型。 本人其它账号:pixiv ID Corey柯瑞,X(twitter)暂时没有,老账号炸了。 图片生成参数及环境:RTX3090 24G,stable diffusion SDXL,noobai底模,模型为Indigo Void Furry Fused XL V2.3,lora是dadadala大佬的,生成时使用adetailer和freeu。采样器为dmp++3m karras,步数65,CFG权重3.0,生成尺寸1024*1024,重绘2倍,重绘迭代40步,修改权重0.6。 欢迎交流探讨,欢迎转载。 Image generated by me, using publicly available models from Civitai. Other accounts: Pixiv ID — Corey柯瑞. No X (Twitter) account for now; my old one got nuked.
Generation setup & environment: RTX 3090 24GB, Stable Diffusion SDXL, base model — noobai, main model — Indigo Void Furry Fused XL V2.3, LoRA by dadadala. Used ADetailer and FreeU during generation. Sampler: DPM++ 3M Karras, steps: 65, CFG: 3.0, resolution: 1024×1024, upscaled 2×, 40 steps for the rework pass, denoising strength: 0.6.
Feel free to chat or exchange ideas, and re-sharing is welcome.