search

JinluZhang1126 / MixSTE

Official implementation of CVPR 2022 paper(MixSTE: Seq2seq Mixed Spatio-Temporal Encoder for 3D Human Pose Estimation in Video)
192 stars 16 forks source link

HumanEva preprocessing error #5

Open xobeiotozi opened 2 years ago

xobeiotozi commented 2 years ago

How did you deal with HumanEva?Or could you share the processed files?

JinluZhang1126 commented 1 year ago

The dataset preprocess follows the VideoPose3D, please check it in https://github.com/facebookresearch/VideoPose3D/blob/main/DATASETS.md.

xobeiotozi commented 1 year ago

If I want to reproduce the HumanEva's results of the paper, what's the command should I have?