
Read the instructions or this won't work well.
This DoRA has an extra thick palette knife /impasto painting style. It does what it does pretty well, but I'm not completely satisfied with it yet, so I'll probably do another one of these in the near future to hopefully iron out the quirks.
Instructions:
-
Use this with Envy Starlight (see below). That's what it was trained on. It may work to varying degrees with other XL checkpoints, but the optimal prompt and strength may differ.
-
Make sure you have this extension installed and enabled:
https://github.com/cheald/sd-webui-loractl -
In the front of your positive prompt, add the following: "oil painting, (shiny globby viscous palette knife on canvas:2.5), bump mapping"
-
In front of your negative prompt, add: "smooth, bokeh, f/1.4"
-
Optionally, add "abstract background" to your negative prompt as well.
-
Use the DoRA with these settings:
<lora:EnvyStarlightThickPaletteKnife01:1.25@0,[email protected],[email protected]:hr=2.5>
You can (and should) experiment with this, obviously, but just be aware that if you just slap the DoRA in there at strength 1 without doing anything special in your prompt, you won't get very good results. As I said, I'll try to make another one of these without these quirks.
描述:
训练词语: oil painting,(shiny globby viscous palette knife on canvas:2.5),bump mapping
名称: envyStarlightThickPalette_01v10.safetensors
大小 (KB): 66128
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success