Config file for Human3.6m results reported in paper

[zzzzzhzh]

Hi, this is a very interesting work, and thanks for making it open-source!
I am wondering if you can provide the config file used for human3.6m dataset in the paper. I noticed a performance difference even after changing the keypoint number in the provided config file from 10 to 16.
Thanks!

[jenjsun]

Thanks for your interest! The config should be the same; could you share more details (ex: about your config file, the environment, or your performance/convergence)? It's been a few years and there could have been underlying changes to pytorch or other packages if the env is not the same.