Spaces:

ali-vilab

/

modelscope-text-to-video-synthesis

like 1.4k

Running on A10G

The human faces are abominations.

hello not all but yes i have writing on the youtube page i have droped all and 2 or 3 are really not good i agree but it is just the bigining of this a "state of the art " nobody have never asked what i have asked before and the big thing is " you can now" bur yes it is not a good result everytimes and i am afraid than if one day the result is perfect like with midjourney we all lose something