This lora is still in development.
The current version of this lora trained on very limited dataset.
This lora aims to provide a robust lora to change any style to realistic photograph.
Prompt:
change the picture 1 to realistic photograph, {description of image}
add more description to the prompt helps the model to identify the content of image and it might provide better transformation.
It might used with anime2real v1
https://civitai.com/models/2110229?modelVersionId=2396073
Recommended strength:
0.75-0.9
Used with anime2real v1
0.75 + 0.75
该lora仍在开发中
目前的版本只在相当有限的训练集中训练
该lora意在提供一个足够泛化的lora,将任意风格图像转换成真实照片,
目前仍有许多不足,可以搭配smile的anime2real v1使用
https://civitai.com/models/2110229?modelVersionId=2396073
推荐强度:
0.75-0.9
搭配anime2real v1使用
0.75 + 0.75