SDXL的动漫基础模型,为了取代NAI的位置。
我花了12GB和一个小时训练。
它目前是目前最适合的动漫基础模型。
因为没有任何外在污染。完全由1.0生成所有数据集。
Comfyui 需要手动输入指定的负面prompt。
负面:
worst quality, low quality, ugly,bad art, bad anatomy,mutation, mutated, extra limbs, extra legs, extra arms,poorly drawn hands, poorly drawn feet, poorly drawn face
SDXL-Anime, XL model for replacing NAI
i just finetune it with 12GB in 1 hour.
it is the Best Basemodel for Anime Lora train.
For no more dataset i use form others,
All dataset generate from SDXL-base-1.0
Comfyui need use negative prompt manually。
negative:
worst quality, low quality, ugly,bad art, bad anatomy,mutation, mutated, extra limbs, extra legs, extra arms,poorly drawn hands, poorly drawn feet, poorly drawn face