
在融合模型的时候,意外得到了这个画风,
尝试用这个模型来生成颜色鲜艳、背景唯美的日系动漫风插图,将会是一个绝妙的选择,
另外,有些2D模型似乎会破坏角色LoRA中的特征,这个模型似乎不太会这样,好耶!
还有,请用高清修复到2倍!越清晰细致的插画自然而然越动人,
和之前上传的FromNewWorld模型不同,我用了另外几个模型以及不同的比例融合了这个模型,但鉴于本人不太会起名字,所以就直接叫它FromNewDream了,原来使用的参数依旧可以用在这个模型中,
最后也请多多支持FromNewWorld吧!您的支持将大大地鼓舞我,
It is a quite cute style I met accidentally which I discover during the model-merging experiments.
If you want to get pictures with bright colors, Japanese-anime-like painting style and dreamlike background sceneries, it can be a great choice to try this base model.
Additionally, I think this model will not disturb the characteristics of many characters when using character LoRAs. I really love this point.
Please use Hires. fix from 512×768 to 1024×1536 or higher. It plays a crucial role to make the picture cleaner and more detailed.
I use different models from those merged in FromNewWorld to merge this model, but I have no talent in naming things. So I have just called it FromNewDream. You can use the same configuration to generate pictures in FromNewWorld.
And please also support FromNewWorld! Your support will be very very inspiring to me.
描述:
训练词语:
名称: ClearVAE_V2.3_fp16.safetensors
大小 (KB): 163411
类型: VAE
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success
名称: fromnewdream_v10.safetensors
大小 (KB): 2082642
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success