文件清单 线路:原始(huggingface.co) · 切换
下载后放进对应的 ComfyUI 目录。models/ 指 ComfyUI 的模型根目录。
说明
Jinnsart Style (MOSS-Image)
作者:Alt4Pai
Model trained on base MOSS-Image (int8) to imitate the style of Jinnsart ===== General note on MOSS-Image: Through the personal testing I've done and what I've seen from others: MOSS is not the ideal candidate for anime style LORAs as of right now, especially compared to a model like Anima. But, it is an extremely powerful model, and I've seen the style LORAs are in demand and do serve some purpose, so I will still gladly train them (especially for some of my Anima favorites). If what I hear of MOSS releasing an anime-centric version of the model someday is true, I'll be retraining a lot of these models in the future. For now, try adding things to the prompt to help describe the art style (e.g. 'glossy anime style illustration' or 'Illustration in a cartoon style with flat bold colors and thick black outlines', etc.), use higher weights, or other more precise prompting methods. NOTE: All the images I generate for the showcase are done with base MOSS-Image (int8_convrot) without the use of any quality tags, negative prompt, Adetailer, Inpainting, etc. only Hires. fix to do very basic upscaling. You can get nicer looking results using MOSS-Image SFT, other finetunes, or just more precise generating methods in general. These outputs on base are to show that the LORA works.
