Skip to content

Latest commit

 

History

History
40 lines (24 loc) · 1.32 KB

README.md

File metadata and controls

40 lines (24 loc) · 1.32 KB

HEXA: Human Demo Augmented Explorer and Achiever

Setup

Create the conda environment by running :

conda env create -f environment.yml

Clone the hexa-benchmark repo, and modify the python path
export PYTHONPATH=<path to lexa-training>/lexa:<path to lexa-benchmark>

Export the following variables for rendering
export MUJOCO_RENDERER=egl; export MUJOCO_GL=egl

WARNING! Make sure to use the right python and mujoco version. The robobin environment code is known to break with other versions. Other environments might or might not work.

Training

First source the environment : source activate lexa

For training, run :

export CUDA_VISIBLE_DEVICES=<gpu_id>  
python train.py --configs defaults lexa_temporal --task kitchen --logdir <log path>

To view the graphs and gifs during training, run tensorboard --logdir <log path>

Acknowledgements

This code was developed on top of LEXA code base https://github.com/orybkin/lexa This code was developed using Dreamer V2 and Plan2Explore.