Merge pull request #37 from Ar-Ray-code/fix/dockerfiles

Fix/dockerfiles

Author: Ar-Ray-code

yolox_ros_cpp/README.md

@@ -7,6 +7,7 @@
 - TensorRT 8.x *
 - ONNXRuntime *
 - Tensorflow Lite *
+- **CUDA 11**
 
 ※ Either one of OpenVINO or TensorRT or ONNXRuntime or Tensorflow Lite is required.
 
@@ -18,106 +19,8 @@
 
 ※ Model convert script is not supported OpenVINO 2022.*
 
-※ YOLOX is not required.
+※ Don't use CUDA 12
 
-※ Jetson + TensorRT docker support (Jetpack 4.6 r32.6.1). Tested with Jetson Nano 4GB.
-
-
-
-
-<!-- <details>
-<summary>Execute with docker</summary>
-
-
-#### OpenVINO
-```bash
-# base image is "openvino/ubuntu20_dev:2021.4.1_20210416"
-docker pull fateshelled/openvino_yolox_ros:latest
-
-xhost +
-docker run --rm -it \
-           --network host \
-           --privileged \
-           --user openvino \
-           -v $HOME/ros2_ws:/home/openvino/ros2_ws \
-           -v /tmp/.X11-unix:/tmp/.X11-unix \
-           -w /home/openvino/ros2_ws \
-           -e DISPLAY=$DISPLAY \
-           --device /dev/video0:/dev/video0 \
-           fateshelled/openvino_yolox_ros:latest /bin/bash
-
-# If use NCS2, mount "/dev/bus/usb".
-xhost +
-docker run --rm -it \
-           --network host \
-           --privileged \
-           --user openvino \
-           -v $HOME/ros2_ws:/home/openvino/ros2_ws \
-           -v /tmp/.X11-unix:/tmp/.X11-unix \
-           -w /home/openvino/ros2_ws \
-           -v /dev/bus/usb:/dev/bus/usb
-           -e DISPLAY=$DISPLAY \
-           --device /dev/video0:/dev/video0 \
-           fateshelled/openvino_yolox_ros:latest \
-           /bin/bash
-
-```
-
-#### TensorRT
-```bash
-# base image is "nvcr.io/nvidia/pytorch:21.09-py3"
-docker pull swiftfile/tensorrt_yolox_ros:latest
-
-xhost +
-docker run --rm -it \
-           --network host \
-           --gpus all \
-           --privileged \
-           -v $HOME/ros2_ws:/root/ros2_ws \
-           -v /tmp/.X11-unix:/tmp/.X11-unix \
-           -w /root/ros2_ws \
-           -e DISPLAY=$DISPLAY \
-           --device /dev/video0:/dev/video0 \
-           swiftfile/tensorrt_yolox_ros:latest \
-           /bin/bash
-```
-
-#### Jetson + TensorRT
-```bash
-# base image is "dustynv/ros:foxy-ros-base-l4t-r32.6.1"
-docker pull fateshelled/jetson_yolox_ros:foxy-ros-base-l4t-r32.6.1
-
-# This image cannot display GUI.
-docker run --rm -it \
-           --network host \
-           --runtime nvidia \
-           -v $HOME/ros2_ws:/root/ros2_ws \
-           -w /root/ros2_ws \
-           --device /dev/video0:/dev/video0 \
-           fateshelled/jetson_yolox_ros:foxy-ros-base-l4t-r32.6.1 \
-           /bin/bash
-```
-
-#### ONNXRuntime
-```bash
-# base image is "nvcr.io/nvidia/cuda:11.4.2-cudnn8-devel-ubuntu20.04"
-docker pull fateshelled/onnxruntime_yolox_ros:latest
-
-xhost +
-docker run --rm -it \
-           --network host \
-           --gpus all \
-           --privileged \
-           -v $HOME/ros2_ws:/root/ros2_ws \
-           -v /tmp/.X11-unix:/tmp/.X11-unix \
-           -w /root/ros2_ws \
-           -e DISPLAY=$DISPLAY \
-           --device /dev/video0:/dev/video0 \
-           fateshelled/onnxruntime_yolox_ros:latest \
-           /bin/bash
-```
-
-</details> -->
 
 ## Clone YOLOX-ROS
 ```bash

yolox_ros_cpp/docker/jetson/README.md

@@ -0,0 +1,33 @@
+# Jetson
+
+## Environment
+
+- Jetson AGX Orin
+- Jetpack 5.1 (R35.1.0 @Ubuntu 20.04)
+- Docker
+- [TIER IV C1 Camera environment](https://github.com/tier4/tier4_automotive_hdr_camera) (When using C1 camera)
+
+## Build & attach shell
+
+
+```bash
+cd ./YOLOX-ROS/yolox_ros_cpp/docker/jetson
+bash setup.bash
+```
+
+## Build (in docker container)
+
+```bash
+source /dependencies/install/setup.bash
+colcon build --cmake-args -DJETSON=ON
+bash src/YOLOX-ROS/weights/onnx/download.bash yolox_tiny
+bash src/YOLOX-ROS/weights/tensorrt/convert.bash yolox_tiny
+```
+
+
+## Run (in docker container)
+
+```bash
+source ./install/setup.bash
+ros2 launch yolox_ros_cpp yolox_tensorrt.launch.py
+```

yolox_ros_cpp/docker/jetson/docker-compose.yaml

@@ -1,65 +1,27 @@
-FROM dustynv/ros:foxy-ros-base-l4t-r32.6.1
+FROM dustynv/ros:humble-ros-base-deepstream-l4t-r35.1.0
 
 ENV DEBIAN_FRONTEND=noninteractive
 
-RUN apt update && \
-    apt install -y libeigen3-dev python3-pip && \
-    apt -y clean && \
-    rm -rf /var/lib/apt/lists/*
-
-RUN python3 -m pip install -U pip && \
-    python3 -m pip install cmake
-
 # add PATH
 ENV PATH=$PATH:/usr/src/tensorrt/bin
 
-# torch 1.9.0
-WORKDIR /workspace
-RUN apt update && \
-    apt install -y libopenblas-base libopenmpi-dev && \
-    apt -y clean && \
-    rm -rf /var/lib/apt/lists/*
-RUN wget https://nvidia.box.com/shared/static/h1z9sw4bb1ybi0rm3tu8qdj8hs05ljbm.whl -O torch-1.9.0-cp36-cp36m-linux_aarch64.whl && \
-    python3 -m pip install Cython && \
-    python3 -m pip install numpy torch-1.9.0-cp36-cp36m-linux_aarch64.whl &&\
-    rm torch-1.9.0-cp36-cp36m-linux_aarch64.whl 
-
-# torch2trt
-WORKDIR /workspace
-RUN git clone https://github.com/NVIDIA-AI-IOT/torch2trt && \
-    cd torch2trt && \
-    python3 setup.py install    
-
-# torchvision
-ENV TORCHVISION_VERSION=0.10.0
-WORKDIR /workspace
-RUN apt update && \
-    apt install -y libjpeg-dev zlib1g-dev libpython3-dev libavcodec-dev libavformat-dev libswscale-dev && \
-    apt -y clean && \
-    rm -rf /var/lib/apt/lists/*
-RUN git clone --branch v${TORCHVISION_VERSION} https://github.com/pytorch/vision torchvision && \
-    cd torchvision && \
-    BUILD_VERSION=$TORCHVISION_VERSION && \
-    python3 setup.py install
-
-# YOLOX
-WORKDIR /workspace
-ENV YOLOX_VERSION=0.2.0
-# matplotlib, onnx-optimizer cannot install by pip
-RUN apt update && \
-    apt install -y python3-matplotlib && \
-    apt -y clean && \
-    rm -rf /var/lib/apt/lists/*
-RUN python3 -m pip install onnx==1.8.1 onnxruntime==1.8.0
-RUN git clone --recursive https://github.com/onnx/optimizer onnxoptimizer && \
-    cd onnxoptimizer && \
-    python3 -m pip install -e .
-
-# bugfix for numpy==1.19.5 
-# set OPENBLAS_CORETYPE environment variable to ARMV8
-ENV OPENBLAS_CORETYPE=ARMV8
-
-RUN git clone -b $YOLOX_VERSION https://github.com/Megvii-BaseDetection/YOLOX YOLOX && \
-    cd YOLOX && \
-    python3 -m pip install -e .
-
+RUN apt-key adv --keyserver keyserver.ubuntu.com --recv-keys 42D5A192B819C5DA && \
+    apt-get update && \
+    apt-get install -y \
+    doxygen \
+    libfmt-dev \
+    librange-v3-dev \
+    python3-pip \
+    python3-jinja2 \
+    python3-typeguard
+RUN mkdir -p /dependencies/src && \
+    cd /dependencies/src && \
+    git clone https://github.com/PickNikRobotics/RSL.git && \
+    git clone https://github.com/PickNikRobotics/cpp_polyfills.git && \
+    git clone https://github.com/ros-perception/vision_opencv.git  -b humble && \
+    git clone https://github.com/PickNikRobotics/generate_parameter_library.git && \
+    cd /dependencies/ && \
+    . /opt/ros/humble/install/setup.sh && \
+    colcon build
+
+COPY /usr/include/aarch64-linux-gnu/ /usr/include/aarch64-linux-gnu/

yolox_ros_cpp/docker/jetson/dockerfile

@@ -0,0 +1,13 @@
+SCRIPT_DIR=$(cd $(dirname $0); pwd)
+cd ${SCRIPT_DIR}
+mkdir ./tensorrt_dir
+cp -r /usr/include/aarch64-linux-gnu/* ./tensorrt_dir/
+
+docker build -t yolox_ros_jetson:latest -f ./dockerfile .
+docker run -it --rm --network host --volume $HOME/ros2_ws:/root/ros2_ws --device /dev/video0:/dev/video0 --workdir /root/ros2_ws --runtime nvidia yolox_ros_jetson /bin/bash
+
+
+# source /dependencies/install/setup.bash
+# colcon build --cmake-args -DJETSON=ON
+# bash src/YOLOX-ROS/weights/onnx/download.bash yolox_tiny
+# bash src/YOLOX-ROS/weights/tensorrt/convert.bash yolox_tiny

yolox_ros_cpp/docker/jetson/setup.bash

@@ -1,11 +1,11 @@
 version: '3.4'
 services:
-    yolox_ros:
+    yolox_ros_onnxruntime:
         container_name: yolox_onnxruntime
-        # build:
-        #     context: .
-        #     args:
-        #         - BASE_TAB=11.4.2-cudnn8-devel-ubuntu20.04
+        build:
+            context: .
+            args:
+                - BASE_TAB=12.1.0-cudnn8-devel-ubuntu22.04
         image: fateshelled/onnxruntime_yolox_ros:latest
         network_mode: host
         runtime: nvidia

yolox_ros_cpp/docker/onnxruntime/docker-compose.yaml

@@ -1,4 +1,4 @@
-ARG BASE_TAG=11.4.2-cudnn8-devel-ubuntu20.04
+ARG BASE_TAG=11.7.1-cudnn8-devel-ubuntu22.04
 FROM nvcr.io/nvidia/cuda:${BASE_TAG}
 ENV	 DEBIAN_FRONTEND=noninteractive
 
@@ -22,12 +22,9 @@ RUN git clone --depth 1 --recursive https://github.com/microsoft/onnxruntime -b
                --use_cuda \
                --config RelWithDebInfo \
                --build_shared_lib \
-               --parallel \
-               --build_wheel \
                --skip_tests && \
     cd build/Linux/RelWithDebInfo && \
     make install && \
-    python3 -m pip install dist/*.whl && \
     rm -r /workdir/onnxruntime
 
 # Install ROS2
@@ -39,41 +36,26 @@ RUN apt update && apt install locales && \
 ENV LANG=en_US.UTF-8
 
 RUN apt update && \
-    apt install -y curl gnupg2 lsb-release && \
-    curl -sSL https://raw.githubusercontent.com/ros/rosdistro/master/ros.key  -o /usr/share/keyrings/ros-archive-keyring.gpg && \
+    apt install -y git wget curl gnupg2 lsb-release && \
+    curl -sSL https://raw.githubusercontent.com/ros/rosdistro/master/ros.key -o /usr/share/keyrings/ros-archive-keyring.gpg && \
     echo "deb [arch=$(dpkg --print-architecture) signed-by=/usr/share/keyrings/ros-archive-keyring.gpg] http://packages.ros.org/ros2/ubuntu $(lsb_release -cs) main" | tee /etc/apt/sources.list.d/ros2.list > /dev/null && \
-    sed -i -e 's/ubuntu .* main/ubuntu focal main/g' /etc/apt/sources.list.d/ros2.list && \
+    sed -i -e 's/ubuntu .* main/ubuntu jammy main/g' /etc/apt/sources.list.d/ros2.list && \
     apt update && \
-    apt install -y ros-foxy-ros-base \
-                   python3-colcon-common-extensions \
-                   ros-foxy-v4l2-camera \
-                   ros-foxy-cv-bridge \
-                   ros-foxy-rqt-graph \
-                   ros-foxy-rqt-image-view && \
-    rm -rf /var/lib/apt/lists/*  &&\
+    apt install -y ros-dev-tools \
+                    ros-humble-cv-bridge \
+                    ros-humble-generate-parameter-library \
+                    ros-humble-parameter-traits \
+                    ros-humble-ros-base \
+                    ros-humble-rqt-image-view \
+                    ros-humble-v4l2-camera && \
     apt -y clean && \
+    rm -rf /var/lib/apt/lists/* && \
     pip install -U pip && \
     pip install catkin_pkg && \
     pip install empy && \
     pip install lark && \
     python3 -m pip cache purge
 
-WORKDIR /workdir
-ENV YOLOX_VERSION=0.3.0
-RUN git clone --depth 1 https://github.com/Megvii-BaseDetection/YOLOX -b $YOLOX_VERSION && \
-    cd YOLOX && \
-    # python3 -m pip install -r requirements.txt && \
-    python3 -m pip install -U numpy \
-                              torch>=1.7 \
-                              opencv_python \
-                              loguru \
-                              tqdm \
-                              torchvision \
-                              thop \
-                              ninja \
-                              tabulate && \
-    python3 -m pip install --no-deps . && \
-    python3 -m pip cache purge
 
 COPY ./ros_entrypoint.sh /ros_entrypoint.sh
 RUN echo "source /ros_entrypoint.sh" >> /root/.bashrc

yolox_ros_cpp/docker/onnxruntime/dockerfile

@@ -1,3 +1,3 @@
 #!/bin/bash
-source /opt/ros/foxy/setup.bash
+source /opt/ros/humble/setup.bash
 source /usr/share/colcon_argcomplete/hook/colcon-argcomplete.bash

yolox_ros_cpp/docker/onnxruntime/ros_entrypoint.sh

@@ -1,15 +1,14 @@
 version: '3.4'
 services:
-    yolox_ros:
+    yolox_ros_openvino:
         container_name: yolox_openvino
-        # build:
-        #     context: .
-        image: fateshelled/openvino_yolox_ros:latest
+        build:
+            context: .
         network_mode: host
         environment:
             - DISPLAY=$DISPLAY
         volumes:
-            - $HOME/ros2_ws:/root/ros2_ws
+            - ../../../:/root/ros2_ws/src
             - /tmp/.X11-unix:/tmp/.X11-unix
         devices:
             - "/dev/video0:/dev/video0"