update docs about YOLOX_ROS_CPP

[swiftfile]

Thanks to this repository, I tried to this node easily!
But, I need extra procedures to run this node completely.
Specifically,When I tried to run yolox_ros(w/docker, tensorRT) following this instruction( yolox_ros_cpp/README.md), I need to install extra dependency not specified in its instruction.

pip install empy
pip install catkin_pkg
pip install lark
apt install ros-foxy-cv-bridge

So I suggest to use my new dockerimage(swiftfile/tensorrt_yolox_ros).

Thank you for all contributors of this repository!
And, I'm glad to create PR for this repo.

[Ar-Ray-code]

Thank you. Your pull requests are welcome!

By the way, do you have any plans to add a Dockerfile?
If possible, it would be nice if you could also list the operating environment.

[swiftfile]

Sorry, my pull request is confusing to see.
If you look at the file diff, you will see the new Dockerfile image name that activated the diff written in the pull request.
And I have changed the instructions to pull my new Dockerfile.

---

My environment is as follows.

$ docker -v
Docker version 20.10.16, build aa7e414
$ uname -a
Linux <myPC> 5.13.0-44-generic #49~20.04.1-Ubuntu SMP Wed May 18 18:44:28 UTC 2022 x86_64 x86_64 x86_64 GNU/Linux
$ apt list |grep nvidia-container

libnvidia-container-dev/bionic 1.9.0-1 amd64
libnvidia-container-tools/bionic,now 1.9.0-1 amd64 [インストール済み、自動]
libnvidia-container1-dbg/bionic 1.9.0-1 amd64
libnvidia-container1/bionic,now 1.9.0-1 amd64 [インストール済み、自動]
nvidia-container-runtime-hook/bionic 1.4.0-1 amd64
nvidia-container-runtime/bionic,now 3.9.0-1 all [インストール済み]
nvidia-container-toolkit/bionic,now 1.9.0-1 amd64 [インストール済み、自動]
$ nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2021 NVIDIA Corporation
Built on Mon_Oct_11_21:27:02_PDT_2021
Cuda compilation tools, release 11.4, V11.4.152
Build cuda_11.4.r11.4/compiler.30521435_0

nvidia-driver version:510.73.08
CPU:i9-12900K
GPU:RTX3090

---

If you would like to know more settings of my operating environment, please tell me.

[Ar-Ray-code]

At this time, I don't know what changes you have made to your Dockerfile.

The pre-built Dockerfile may not work with some versions of the base NVIDIA driver because the CUDA version is fixed.
What if a user receiving the error would be confused by the unmodified yolox-ros-cpp/Dockerfile.

Please add a new Dockerfile or update your existing Dockerfile.
Changes only in the README without source code are unacceptable.

[swiftfile]

Sorry for the late reply.
I have updated Dockerfile.
I tried to apply the changes to "Docker-compose.yaml" as well, but I could not get confirmation that it works.
I would like to get a working confirmation including docker-compose file. However, I don't have enough time now and it is difficult to get a working check.

[Ar-Ray-code]

I followed the procedure and confirmed that it builds Dockerfile and yolox_cpp (tensorrt) correctly.
Thanks!

[swiftfile]

Thanks to check！ ❤️