You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi folks, we're experiencing some problems with MocapNet.
Every video we run through MocapNet winds up generating random noisy animation. In all of these animations, the feet are pulled behind the body, and the arms and legs randomly flail around (see attached).
We're observing this on both versions of the model, and we're observing this with and without post-processing steps applied. We've even run this on the provided shuffle sample and get the same results.
Are we using the incorrect code, checkpoints, arguments, and/or post-processing steps? Can you direct us to the correct configuration?
Dear @echelon , sorry for the very delayed response.
Last weeks have been crazy. I had to travel for a new project and previous Wednesday I successfully defended my PhD thesis which I now have to make the final changes to its text.
That being said there are multiple reasons why tracking problems may occur, I think the best way to help you would be if you send me the video file you want to capture as well as the steps you follow so I can try to identify the problem.
You can use this issue to point to a link to the data if you think this might also help others. If for reasons of confidentiality you have a problem sending me actual data, please attempt to record a similar scenario.
and mention MocapNET on the subject because I get a ton of spam and it is really difficult to keep track of everything!
Looking forward to your input!
And hope I can help you resolve the issue!
Hi folks, we're experiencing some problems with MocapNet.
Every video we run through MocapNet winds up generating random noisy animation. In all of these animations, the feet are pulled behind the body, and the arms and legs randomly flail around (see attached).
We're observing this on both versions of the model, and we're observing this with and without post-processing steps applied. We've even run this on the provided shuffle sample and get the same results.
Are we using the incorrect code, checkpoints, arguments, and/or post-processing steps? Can you direct us to the correct configuration?
I think it may be related to this ticket.
I'd be happy to pay a consulting fee (~$1000) for this.
Thanks so much!
The text was updated successfully, but these errors were encountered: