ROS 2 support of OpenMMLab projects using TensorRT.
- Ubuntu 22.04
- ROS 2 Humble
- TensorRT>=8.6
-
Install ROS 2. Please refer to OFFICIAL DOCUMENT of ROS 2 HUMBLE.
-
Install CUDA/CUDNN/TensorRT. Please refer to OFFICIAL DOCUMENT of NVIDIA.
-
Clone repository and install dependencies:
git clone [email protected]:ktro2828/mmros && cd mmros rosdep update rosdep install -y --from-paths . --ignore-src --rosdistro $ROS_DISTRO
-
Build all packages
colcon build --symlink-install --cmake-args -DCMAKE_BUILD_TYPE=Release
-
(OPTIONAL) Build Only Specific Project
Note that
<PROJECT_NAME>
represents the name of the particular package under theprojects
directory.colcon build --symlink-install --cmake-args -DCMAKE_BUILD_TYPE=Release --packages-up-to <PROJECT_NAME>
Task | Modality | Model |
---|---|---|
Detection | Camera | YOLOX (CVPR'2021) |
Camera | RTMDet (ArXiv'2022) | |
Semantic Segmentation | Camera | Mask2Former (CVPR'2022) |
Camera | PIDNet (ArXiv'2022) | |
Instance Segmentation | Camera | Instance RTMDet (ArXiv'2022) |
Panoptic Segmentation | Camera | Panoptic FPN (CVPR'2019) |
We are welcome your contribution!!
Before starting your work, please follow the Contributing Guidelines.