想法是好的,但是鉴于扩散模型的问题,这个lora变成了一个失败的尝试。
尽管你确实可以用这个lora去生成具有很大角度透视的图像,但是手的图像往往是畸形的(我会在例图中展示一些)。我可能在几天之后,对这个lora进行更新,使其得到一定的改善
The idea is good, but given the problems with the diffusion model, this lora turns out to be a failed attempt.
Although you can indeed use this lora to generate images with a large angular perspective, the hand images tend to be distorted (I'll show some of this in the example image). I may update this lora in a few days to improve it somewhat
Description
FAQ
Comments (7)
Maybe you should keep your dataset consistent with all hands
Thanks!
Now make a POV punching LORA ^_^
yes pls
这个就像使用outstretched arm的prompt去出图, 想着手伸出画面就好了, 结果常常画面上有手, 还是要抽卡
非常期待
我觉得可以用了 毕竟扩散模型 手的话拿controlNet控制或者自己手动画应该不错
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.



