u-net layered fusion, mainly for 3d effects, can be characters, animals, clothing, various images, scenes, backgrounds, landscapes, etc.

The example styles are quite diverse, everyone can take a look, and also be described in natural language

The base film mainly pursues generalization in various aspects, of course, loa's generalization is also good,

Also, thank you to all the model authors for having so many good models,

Recommend high-definition enlargement, weights are arbitrary

Sampling recommendation DPM++SDE, steps 20, CFG 9 or less, recommended width and height 512,680; 640,512, for scene type images recommend 1024 and above

The enlargement algorithms I often use are 4x-ultrasharp and 8x_nmkd-superscale_150000_g

Dynamic thresholding can be used, or CFG can be tried directly from low to high, there are different surprises

Remember to use vae, otherwise the picture may be gray, not recommended 0 for hires steps~~

v2.0 mainly enhances the 3d effect, making it more three-dimensional, and enhances the effect of light and shadow

In addition, the probability of hand collapse is also greatly reduced, all the hands in the example images are one-time outputs, without any post-editing

The prompt words used in the v2.0 version example images are exactly the same as those in the v1.0 version, you can compare the two and use them as appropriate

Generalization is still quite good

For facial repair, please use adetailer.

https://github.com/bing-su/adetailer

You can try enabling dynamic thresholding

https://github.com/mcmonkeyprojects/sd-dynamic-thresholding

Lora layered reference:

github - hako-mikan/sd-webui-lora-block-weight