More envs are on the way...
task | action space | visualize |
---|---|---|
reach | xyz | |
push | xyz | |
pegbox | xyz | |
lift | xyzw | |
shelfplacing | xyzw | |
peginsert | xyzw |
create a newconda env (or just install these packages included):
conda env create -f environment.yaml
And then you could install xArm package by:
python setup.py develop
See example.py
for simple usage.
- observation_type:
state
,image
,state+image
. Note thatstate+image
means using the robot state (xyzw). - domain_randomization: whether to use domain_randomization