Updated the docs for benchmark

This commit is contained in:
felipecode 2017-12-21 21:11:12 -02:00
parent 7b7375e3e2
commit 3329471b11
1 changed files with 10 additions and 13 deletions

View File

@ -9,18 +9,15 @@ used in the CoRL 2017 paper. By running this benchmark you can compare
the results of your agent to the results obtained by the agents
show in the paper.
The basic functionality requires only the protobuf module to be installed
$ sudo apt-get install python3 python3-pip
$ sudo pip3 install protobuf
However, other operations as handling images require some extra modules, and the
"manual_control.py" example requires pygame
$ sudo pip3 install numpy Pillow pygame
The script "PythonClient/client_example.py" provides basic functionality for
controlling the vehicle and saving images to disk. Run the help command to see
options available
$ ./client_example.py --help
The script "PythonClient/manual_control.py" launches a PyGame window with
several views and allows to control the vehicle using the WASD keys.
$ ./manual_control.py --help
! Currently not tested on python 3
Besides the requirements of the CARLA client,
the benchmark package also needs the future package
$ sudo pip install future
Run the help command to see options available
$ ./run_benchmark --help