
I created a mix of models for SDXL - recipe stored in a notebook somewhere and have been using it for a while. I can only store 1 model on my cloud drive and I wanted a model that can do photorealism, nsfw, and artwork at the same time. I wanted my model to not suggest things on its own and rather follow my prompts.
The benefits of this model should be:
-
focussed on photorealism, portraits, and slightly more nsfw.
-
better quality images without distorting the idea/concept in the prompt
-
styles should be preserved. Unlike many photorealism models I see, your artwork should not be converted into a photo with the current model.
Please, check existing images to see how the model closely follows provided prompts.
Things to note:
-
Refiner is not needed.
-
You need to tell exactly what you want in the prompt. Do not expect that prompts from other models/images will produce a great replica. This model, IMO, does approx. what its told in the prompt. You need fire in your image, specify it!
-
I have included some examples from popular images on Civit - you can compare them with their original version to see the difference.
描述:
训练词语:
名称: sahastrakotiXL_v10.safetensors
大小 (KB): 6775431
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success
