search

chaneyddtt / Generating-Multiple-Hypotheses-for-3D-Human-Pose-Estimation-with-Mixture-Density-Network

Code for our CVPR2019 paper: Generating Multiple Hypotheses for 3D Human Pose Estimation with Mixture Density Network
MIT License
97 stars 9 forks source link

About reproducing the result in table 1 #16

Closed 908760230 closed 3 years ago

908760230 commented 3 years ago

I downloaded the two pretrianed model for testing, but i only got the result of Table 6! I want to reproduce the result of table 1.

Could you tell me how should i do ?

And I have some questions as follow: 1) the test dataset in table 1 is ground truth or stackhour dataset ? 2) Creating 2 bi-layers of 1024 units. Creating model with fresh parameters. Working on epoch 1, batch 100 / 24371... done in 69.00 ms Working on epoch 1, batch 200 / 24371... done in 59.07 ms Working on epoch 1, batch 300 / 24371... done in 59.58 ms Working on epoch 1, batch 400 / 24371... done in 59.87 ms Working on epoch 1, batch 500 / 24371... done in 59.36 ms Working on epoch 1, batch 600 / 24371... done in 60.87 ms Working on epoch 1, batch 700 / 24371... done in 58.83 ms Working on epoch 1, batch 800 / 24371... done in 59.55 ms Working on epoch 1, batch 900 / 24371... done in 59.67 ms envirnment : one rtx2080 ti GPU card the training speed is so slow! do you have the same training time ?