Skip to content

Latest commit

 

History

History
executable file
·
132 lines (98 loc) · 4.2 KB

README.md

File metadata and controls

executable file
·
132 lines (98 loc) · 4.2 KB

RangeNetTrt8 for ROS2 Humble

License

Deploy RangeNetTrt8 on ROS2

Prerequisites

1. libtorch

tested on CUDA 11.7, libtorch 2.0.1

wget -c https://download.pytorch.org/libtorch/cu117/libtorch-cxx11-abi-shared-with-deps-2.0.1%2Bcu117.zip
$ unzip libtorch.zip

2. TensorRT, CUDA, cuDNN

Ubuntu GPU TensorRT CUDA cuDNN —
20.04 TITAN RTX 8.2.3 CUDA 11.4.r11.4 cuDNN 8.2.4 ✔️
20.04 NVIDIA GeForce RTX 3060 8.4.1.5 CUDA 11.3.r11.3 cuDNN 8.0.5 ✔️
20.04 (test) NVIDIA GeForce RTX 3070Ti 8.4.3.1 CUDA 11.7.r11.7 cuDNN 8.9.4.25 ✔️
22.04 NVIDIA GeForce RTX 3060 8.2.5.1 CUDA 11.3.r11.3 cuDNN 8.8.0 ✔️
22.04 NVIDIA GeForce RTX 3060 8.4.1.5 CUDA 11.3.r11.3 cuDNN 8.8.0 ✔️

Note

Tensorrt > 8.4 may have some running errors

gedit ~/.bashrc

# CUDA
CUDA_PATH=/usr/local/cuda/bin
CUDA_LIB_PATH=/usr/local/cuda/lib64

# TensorRT
export LD_LIBRARY_PATH=${HOME}/TensorRT-8.4.3.1/lib:$LD_LIBRARY_PATH
export LIBRARY_PATH=${HOME}/TensorRT-8.4.3.1/lib::$LIBRARY_PATH

# libtorch
export Torch_DIR=${HOME}/libtorch/share/cmake/Torch
export LD_LIBRARY_PATH=${HOME}/libtorch/lib:$LD_LIBRARY_PATH

3. ROS2(Humble)

Note

Debian packages for ROS 2 Humble Hawksbill are currently available for Ubuntu Jammy (22.04).
Source installation is available for Ubuntu Focal (20.04).

4. other dependencies

sudo apt install build-essential python3-dev python3-pip apt-utils git cmake libboost-all-dev libyaml-cpp-dev libopencv-dev python3-empy
pip install catkin_tools trollius numpy

Install

# mkdir -p ~/ros2_ws/src && cd ~/ros2_ws/src
git clone https://github.com/NeSC-IV/RangeNetTrt8
git clone https://github.com/ros2-gbp/pcl_msgs-release
cd pcl_msgs-release
git checkout release/humble/pcl_msgs
cd ..
git clone https://github.com/ros-perception/perception_pcl.git
cd perception_pcl
git checkout ros2
cd ../.. && colcon build --symlink-install

Pretrained model github | 百度云

data structure

.
├── RangeNetTrt8
│   ├── model
│   │   ├── arch_cfg.yaml
│   │   ├── data_cfg.yaml
│   │   └── model.onnx

Usage

The first run takes a while to generate the TensorRT optimization engine

img 

# >>> ROS2 >>>
cd ~/ros2_ws/
colcon build --symlink-install
source install/setup.bash
ros2 launch rangenet_pp ros2_rangenet.launch
ros2 launch rangenet_pp ros2_bag.launch

FAQ

❓ 问题 1： [libprotobuf ERROR google/protobuf/text_format.cc:298] Error parsing text-format onnx2trt_onnx.ModelProto: 1:1:

1）情况一：下载的 ONNX 模型不完整，模型解析出问题。重新下载即可。

❓ 问题 2： TensorRT 从 8.2 升级到 8.4 时，预测结果异常，详见 issue#8

不对 235 层的模型权重不进行优化即可

❓ 问题 3： error: A __device__ variable cannot be marked constexpr CUDA 版本过低，需要升级 CUDA 版本（详见 [issue#4](https://github.com/Natsu-Akatsuki/RangeNetTrt8/issues/4)），若要保持较低的版本如 CUDA 11.1 则详见 [issue#2](Natsu-Akatsuki#2)