A spatial LoRA model created based on the ChilloutMix for realistic rendering.
The underlying image dataset comes from design diagrams of certain tea brands, which were used for training. This model can generate visual concepts of spaces from text prompts, as well as enable image-to-image transformations and texture modifications with CN control.
For high-definition restoration, it is recommended to set the redraw scale to around 0.5-0.6.
Weights: A range of 0.6 to 0.8 will yield better results.
Sampling methods: DPM++2M Karras or Euler
Number of iterations: 20
一个基于真实感必备模型｜ChilloutMix制作的空间类lora。
底层的图集来源于部分茶饮品牌的空间设计图，以此进行的训练。可以做一些文生图出空间的意向，也可以用图生图以及CN控制改变材质等等。
高清修复建议开到重绘0.5-0.6左右。
权重：0.6~0.8，出图效果会更好一点
采样方法：DPM++2M Karras或者Euler
迭代步数：20
Description
A spatial LoRA model created based on the ChilloutMix for realistic rendering.
The underlying image dataset comes from design diagrams of certain tea brands, which were used for training. This model can generate visual concepts of spaces from text prompts, as well as enable image-to-image transformations and texture modifications with CN control.
For high-definition restoration, it is recommended to set the redraw scale to around 0.5-0.6.
Weights: A range of 0.6 to 0.8 will yield better results.
Sampling methods: DPM++2M Karras or Euler
Number of iterations: 20
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.