Update:I make the V3 version which differs from the previous version in that the art style will now also affect the characters. In the past versions, for complex prompts, the style of the characters was still based on the base model's style, and only the background was affected by Lora. In this new version, I used the wd14-tagge to tag the captions, which has a greater global impact on the model. (Perhaps this has a greater impact on the base NAI model.)

This time I used LoCon instead of LoHa to make this model, as it seems that LoHa's performance is not very good.

This is the conclusion: In theory, if using stable diffusion 1.5, V2 might be better, but you need to add "a painting of" in front of the prompt. If you don't want this oil painting style to affect the characters, you can also download the V2 version. The V3 version only makes the style more influential on the characters, but this style may not be acceptable to everyone.

The following comparison image shows the difference between V2 and V3 on the effect of styles on people. (The image on the left is V3, and the image on the right is V2.)

Note:

This is a Lycoris model, and you need to install a1111-sd-webui-locon extension to use this model.

And don't use sd-webui-additional-networks to load this model.

Place the model in the path '.\stable-diffusion-webui\models\Lora' to use it.

Suggest weight:around 1 or higher.(But not suggest too higher,Of course, you can adjust it according to your needs.)

Optional tag: traditional media. (may enhance the effect)

Suggest weight: around 1 Or higher.(like 2,if you want)

Suggest CFG scale:1.5-3

It is not guaranteed to work well with other LoRA , The art style may not affect the LoRa characters.(If that's what you want) .It is recommended to increase the weight and use low CFG scale.

Based on novel-anime-full model.