Distilling and retrieving generalizable knowledge from online human language corrections.
Create the environment and start using the repo.
conda env create -f environment.ymlpython scripts/script.py --realrobot <realrobot> --task <task>where <realrobot> is a bool variable that determines whether this experiment is run on a real robot or not, <task> is the name of the task to run. You have to pre-define the available tasks along with their clip candidates in utils/perception/perception_utils.py.