V3.0 verson更好的还原度，更好的人体，更温馨的画风better restoration, better human figures, and a warmer art style.Training model: Animagine XL V3.1Inference model: Animagine XL V3.1Training resolution: 1024*1024Recommended output resolution: 1024*1280 /832*1216Recommended model weight: 0.8~1-24.4.10更新：Animagine XL V3.1训练再次感谢kozue佬提供的训练集。24.1.11更新：该模型是在C站发布的第一个基于Animagine XL V3的画风lora。24.3.4更新：该模型是在C站发布的第一个基于Kohaku-XL Delta的画风lora-Delta版本更新 Delta version updateBased on <a target="_blank" rel="ugc" href="https://civarchive.com/models/332076">Kohaku-XL Delta - rev1 | Stable Diffusion Checkpoint | Civitai</a> trainingTraining model: Kohaku-XL DeltaInference model: Kohaku-XL DeltaTraining resolution: 1024*1024Recommended output resolution: 1024*1280Recommended model weight: 0.8~1_2.0版本更新 2.0 version updateBased on <a target="_blank" rel="ugc" href="https://civarchive.com/models/260267/animagine-xl-v3">https://civarchive.com/models/260267/animagine-xl-v3</a> trainingTraining model: Animagine XL V3Inference model: Animagine XL V3Training resolution: 1024*1024Recommended output resolution: 1024*1280Recommended model weight: 0.8~1基于<a target="_blank" rel="ugc" href="https://civarchive.com/models/260267/animagine-xl-v3">https://civarchive.com/models/260267/animagine-xl-v3</a> 训练训练模型：Animagine XL V3跑图模型：Animagine XL V3训练分辨率：1024*1024建议出图分辨率：1024*1280建议模型权重：0.8~11.0版本更新：更像的画风，更好的光影，但是代价是，生成图片的肢体似乎会受到影响。（？）Version 1.0 Update:A more realistic art style and improved lighting effects have been implemented. However, the trade-off is that the generated images may be affected in terms of their body proportions.（？）这是一个画风模型，是对《COMIC LO》这一杂志画风的模仿，他的画风给人一种清新舒服的感觉。我尊重人类画师，并不认为ai可以否定人类画师的创作，本模型仅做展示交流，使用本模型以及创作出的图片不得用于商业用途。感谢kozue佬对我的指导，这个模型名称的灵感也是源于他。训练模型：SDXL1.0官模测试模型：Kohaku-XL该模型不需要任何触发词。This is an art style model that imitates the art style of the magazine "COMIC LO." Its art style gives a refreshing and comfortable feeling. I respect human artists and do not believe that AI can negate the creativity of human artists. This model is only for demonstration and communication purposes. The use of this model and the created images is not permitted for commercial purposes.Thank you to my guide, Kozue, for their guidance. The inspiration for the name of this model also came from them. (Note: "佬" is a slang term commonly used in some Chinese dialects, particularly Cantonese, to refer to someone respectfully or playfully.)Training model: SDXL1.0 Official ModelTesting model: Kohaku-XLThis model does not require any trigger words.

C萝画风  XL comic lo takamichi style XL

col_xl.safetensors

Files with SHA256: c5fec75e232d692fd7bcfb3a19ef5c60f9c6d687265f0c0356f88ea1e309b583