Run a PyTorch model in C++ in ROS2. A template node to be able to adapt to whatever model you want to run, for instance object detection of people and bikes.
There might be major changes to the code as it is a WIP.
/ - topic for
Make sure to run setup.bash and local_setup.bash for your ROS2 environment for all dependencies or symlink them into the repo.
Add Libtorch to your CMAKE_PREFIX_PATH env variable, like so:
export CMAKE_PREFIX_PATH=$CMAKE_PREFIX_PATH:/path/to/libtorch
Add CV_Bridge as well if its not already on your CMAKE_PREFIX_PATH
export CMAKE_PREFIX_PATH=$CMAKE_PREFIX_PATH:<path>
(If its built using colcon build this is the path to cv_bridge folder in the install folder.)
Run
colcon build
Unfortunately you also have to copy two libs from https://github.com/intel/mkl-dnn/releases into the libtorch lib folder.
libmklml.dylib
libiomp5.dylib
ros2 run ros2_pytorch ros2_pytorch_node