大神进来 视频转绘问题

目前可以i2v, 或者t2v+pose, 但是好像不能用i2v+pose的方式??

比如:
i2v(也可以promt控制), 加上pose控制(或者controlnet),
我的目的就是为了实现和animatediff差不多的功能, 虽然andimatediff可以加一些放大或者controlent之类的生成, 但是animatediff生成的画面很多噪点还有画面不一致. 有大神指教一下吗??

测试了cogvideox-fun i2v+pose ,速度比animatediff lcm还要快. 效果也好很多....

Currently i2v, or t2v+pose can be used, but it seems that i2v+pose cannot be used.

For example:
i2v(also promt control), plus pose control (or controlnet),
My purpose is to achieve similar functions to animatediff, although andimatediff can add some magnification or controlent generation, but animatediff generated a lot of noise and screen inconsistency. Do you have some advice from the great God?

The cogvideox-fun i2v+pose was tested even faster than the animatediff lcm. It works a lot better....

The issue does not identify files, tests, or entry points. First inspect how VideoX-Fun currently handles i2v, t2v, pose, and ControlNet workflows; done would mean a documented or working i2v-plus-pose path comparable to the requested CogVideoX-Fun behavior.