I am very interested in the model and would like to give it a try. I have some questions about model training. It would be great if you could give me some instructions on them.
(1) I have some short video pieces of the same person. Each is about 20 seconds long with pose differences. Would it work if I stitch all these videos one by one to form a long video and train the model?
(2) If I want to train on my own video, would it work if the person's face in the video is not frontal (i.e. the person is not looking at the camera)?
(3) Do I need to re-train SyncNet and Audio2motion if I want to use mandarin audio to drive a face like this demo?
Hi,
I am very interested in the model and would like to give it a try. I have some questions about model training. It would be great if you could give me some instructions on them.
(1) I have some short video pieces of the same person. Each is about 20 seconds long with pose differences. Would it work if I stitch all these videos one by one to form a long video and train the model?
(2) If I want to train on my own video, would it work if the person's face in the video is not frontal (i.e. the person is not looking at the camera)?
(3) Do I need to re-train SyncNet and Audio2motion if I want to use mandarin audio to drive a face like this demo?
Thanks for your help!