This LyCORIS model focuses on the style of thick oil painting, suitable for Stable Diffusion 1.5. I recommend using Dreamshaper 8 in conjunction.

If you are using the webUI of automatic1111, you must place the model in the models/LyCORIS path before using it.

The activation keyword for this model is: rick_thickoil. A typical prompt structure is as follows:

rick_thickoil, a painting of ____, <lora:Rick_ThickOil:0.8>

In most cases, setting the weight to 0.8 can achieve good results. If the weight is lowered, for example to 0.3, the effect of thick application will weaken, like a regular illustration.

The content of the image also affects the effect of thick coating. Landscapes will have the strongest thick coating effect, followed by buildings, and finally, people: Landscapes > Buildings > People.

I often use this model to convert the photos I take with my phone into oil paintings. Here is a reference workflow:

1. Use img2img to reference the photo.

2. Enable two controlnets, one for Depth and one for Canny, with a weight of 1 for both.

3. Adjust the Denoising strength starting from 0.65.

4. Other parameters are similar to txt2img.

这是一个专注于厚涂油画风格的 LyCORIS 模型，适用于 Stable Diffusion 1.5. 我推荐使用Dreamshaper 8来配合使用。

如果你使用 automatic1111 的 webUI，使用前需要把模型放在 models/LyCORIS 路径下。

本模型的激活词是：rick_thickoil，一个典型的提示词结构如下：

rick_thickoil, a painting of ____, <lora:Rick_ThickOil:0.8>

在大多数情况下，权重设置为0.8可以得到不错的效果，如过把权重降低，比如0.3，厚涂的效果会变得很弱，就像普通的插画了。

画面的内容对厚涂效果也有影响，风景会得到最强的厚涂效果，其次是建筑，最后是人物：风景>建筑>人物。

我常常用这个模型来把我用手机拍的相片转为油画，这是一个参考的工作流：

1、使用 img2img，把相片作为参考。

2、启用2个controlnet，一个是Depth，一个是Canny，权重都为1.

3、Denoising strength 可以从0.65开始进行微调。

4、其他参数跟 txt2img 类似。

希望你喜欢。我的小红书帐号：王小王，分享AI技术和知识，欢迎交流。