Julieta Martinez

@Seleucia Thanks for finding this. After going through my code again I believe that this is indeed a bug. However, the reason and its impact are a little more subtle...

*Heads up*: Very jetlagged me writing this post. Remember, in the paper we proposed using a "sampling-based" loss, which we found produces (relatively) realistic motion in the long-term and does...

^ The length of sampling in an RNN does not affect the results (other than showing you more samples).

^^ I believe so. If you wish to compare against this paper, compare against the output using `--sample`, and against the suprisingly-hard-to-beat zero velocity baseline.

Hi! I think this has been reported before: https://github.com/una-dinosauria/3d-pose-baseline/issues/66 Please feel free to submit a PR if you know the fix. Cheers,

Hi @jutanke, Nice catch. There seems to be a reflection about the x axis. This seems quite strange to me to be honest, as it's hard for me to image...

Hi @jutanke, Thanks for your great work tracking down this issue, and sorry for the delay in responding; I was very busy with a deadline. I suspect this is probably...

Yes, the first three numbers are global position and the next three are global orientation. Sorry for taking long to answer. Deadlines!

Hi, Sorry this is pretty hard for me to follow. Could you point to me to specific lines in the code, and what you think they should be instead? Better...

I think I see it. It works fine when there is only one action, but not when we train on all actions. Would you be up for submitting a PR?