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
In the article corresponding to this repository, in section 7.4 Acyclic Motions, is is mentioned an idea about additionally conditioning the VAE model. It is said that it can generate the next frame considering a one-hot encoding of an acyclic action like kick or header.
I have gone through the repository code, but I might have missed this part. As far as I have seen, at least for the default model (Mixture VAE), the encoder only conditions on the given number of previous frames (https://github.com/electronicarts/character-motion-vaes/blob/main/vae_motion/models.py#L87). Did I miss this additional conditioning part or it was not included in current state of the repository?
Thank you!
The text was updated successfully, but these errors were encountered:
You’re right — you haven’t missed it, it’s not in the current state of the repository. The acyclic motion experiments were late additions to the paper (suggested by reviewers), and so didn’t make it to the code release.
We chose the easiest implementation that we could come up. In L87 of the encoder, the frame_size would be larger, as it would contain the one-hot vector as well as the pose. By treating the one-hot vector as part of the condition, we can implement it with minimal code changes. The decoder can be similarly modified. When training RL controller or at test time, we would force the condition vector to be a certain action by blending the one-hot vector with the desired action.
Hello,
In the article corresponding to this repository, in section
7.4 Acyclic Motions
, is is mentioned an idea about additionally conditioning the VAE model. It is said that it can generate the next frame considering a one-hot encoding of anacyclic
action likekick
orheader
.I have gone through the repository code, but I might have missed this part. As far as I have seen, at least for the default model (Mixture VAE), the encoder only conditions on the given number of previous frames (https://github.com/electronicarts/character-motion-vaes/blob/main/vae_motion/models.py#L87). Did I miss this additional conditioning part or it was not included in current state of the repository?
Thank you!
The text was updated successfully, but these errors were encountered: