Official code for the ICCV 2021 paper "DECA: Deep viewpoint-Equivariant human pose estimation using Capsule Autoencoders". All the code is written using Pytorch Lightning. Please use Pipenv to configure the virtual environment required to run the code.
Use the following command to configure the virtual environment:
pipenv install
To configure all the network parameters, including the dataset paths and hyperparameters, please edit the file:
config/config_TV.cfg
or add each parameter as a runtime flag while executing the main.py file as follows:
python main.py --flagfile config/config_TV.cfg
As an example, to run the network in training mode with a dataset stored in , you can run the following command:
python main.py --flagfile config/config_TV.cfg --mode train --dataset_dir <datasetpath>