Getting Started
Here is a quick example of how to train and run a teacher environment:
(.venv) $ highrl
The teacher will start training and will output the results (traning graphs for both teacher and robot on the Desktop).
You can view the command-line args:
(.venv) $ highrl -h
USAGE: Parse arguments [-h] [--robot-config ROBOT_CONFIG_PATH] [--teacher-config TEACHER_CONFIG_PATH] [--mode {train,test}] [--env-mode {teacher,robot}] [--render-each RENDER_EACH] [--output-dir OUTPUT_DIR] [--lidar-mode {flat,rings}]
Here is the full of command-line args:
- -h, --help
show help message and exit
- --robot-config
path of configurations file for robot env
- --teacher-config
path of configurations file for teacher env
- --mode=train
mode of operation
- --env-mode=teacher
which env to use through training/testing
- --render-each
frequency of rendering for robot environment
- --output-dir
relative path to output results for robot mode
- --lidar-mode=flat
mode to process lidar flat=1D, rings=2D
Note
The library is built with default configurations, however, you can change those configurations by incorporating your own .config files. See section configs for more details.