这个模型是碰巧做出来的，下面是我自己尝试的用法：1、2图（链条主要在身体上）：先用real类模型出一张链条比较合适的图（我用了UberRealisticPornMerge），身体比例没有太大问题即可，然后转到图生图，切换模型（我用了anything v4.5）再画一张（建议重绘幅度6~7），此时建议也启用此lora模型。3、4图（到处都有链条）：正常使用即可，但会受tag影响，比如使用outdoors、streets等tag可能会大大减少环境中的链条（如图5，但腹部链条也往往较少）。可能仍有我没发现的东西，欢迎在评论区分享。Pictures 1、2(chains are mainly on the body): use a real model to create a picture with some satisfactory chains and a normal body(I used UberRealisticPornMerge), then use img2img and switch the model (I used anything v4.5), Draw a new picture(recommended redraw range 6~7), it is recommended to enable this lora model at this time.Pictures 3、4(chains everywhere): just use the model normally. but it will be affected by tags. For example, using tags like outdoors、streets may greatly reduce the chains in the environment(like picture 5, but chains on belly tend to be less).There may still be something I haven't found, hope you can share in the comment below.
