Help fuel my passion! Even $1 makes a difference.
这是一个全新的精调训练lora模型，使用超过500张高质量图片进行训练。
训练图片全部来自该画师的原画，得到纯净的村田莲尔风格LORA。
碎碎念：
其实村田莲尔的lora社区里已经蛮多的了，效果也还不错，
因为一直有人希望我做一个，那么就顺手丢一个吧~~就是这样，嗯！╮(╯▽╰)╭
推荐权重：0.8~1.5
--- --- --- --- --- --- --- ---
This is a brand new fine-tuned training lora model, which is trained with more than 500 high-quality pictures.
All the training pictures are from the original painting collection of the artist, so that the pure Artist style lora is obtained.
Recommendation Weight：0.8~1.5
Description
FAQ
Comments (5)
Good lora, it captures the artist style perfectly but offen creates grey and white borders. Most likely due to training photos has borders. I myself have made the same artist style. I cut the borders before traning. Maybe a second version would make this lora perfect.
In fact, in order to better achieve the original effect of the artist, I deliberately kept some of the training pictures with borders.
In general, it will appear in the addition prompt: traditional media;
If you don't want borders, you can add the negative prompt word: border, and remove the prompt traditional media.
Awesome style!