文件清单 线路:镜像(hf-mirror.com) · 切换
下载后放进对应的 ComfyUI 目录。models/ 指 ComfyUI 的模型根目录。
说明
artanis69 style | Anima LoRA
作者:mnemosynekotorin
Art-style LoRA. Trained on Anima-Base v1.0. Use at 0.7-0.9 strength (0.8 in the previews). Trigger: artanis69, put it first after the quality tags. Works with any character or OC. Recommended: Euler a, 24-30 steps, CFG 4-5, 832x1216 (hires 1.5x, denoise 0.35). Previews made on Hoseki LustrousMix [Anima v1.0]; works on Anima-Base v1.0 too. Prompt start: masterpiece, best quality, newest, highres, absurdres, 1girl, solo, Negative: low quality, worst quality, score_2, score_1, watermark, bad anatomy, signature, jpeg artifacts, logo, bad hands, artist name, missing finger, extra digits, fewer digits, disfigured, mutation, monochrome, greyscale, loli, shota, chibi, text, patreon username, twitter username, web address (keeps stray signatures out) Anima-based model, released under the CircleStone Labs Non-Commercial License.
