Ubuntu 18.04
📌설치할 때 Normal 말고 Minimal installation 선택

---

1. Nvidia Driver 설치

• Software & Updates > Additional Drivers 에서 아래 드라이버 선택

  Using NVIDA driver metapackage from nvidia-driver-510 (proprietary)

• Apply Changes 누르고 Restart
• 터미널에서 nvidia-smi 명령어 입력해서 잘 출력되는지 확인하기

---

2. Anaconda 설치

Anaconda | Anaconda Distribution

• 설치파일: https://repo.anaconda.com/archive/Anaconda3-2022.05-Windows-x86_64.exe 다운로드
• 아래 명령어 입력

  bash ~/Downloads/Anaconda3-2022.05-Windows-x86_64.sh

• 터미널 끄고 새로 열면 아나콘다 반영 완료

---

3. CUDA 11.3 설치

CUDA Toolkit 11.3 Update 1 Downloads

📌 runfile로 설치하기

📌 꼭 Driver 선택해제!

• (안하면 드라이버 충돌나서 nvidia-smi 이상해짐)

  • 아래와 같은 오류 생김

    nvidia-smi 실행 시, Driver/library version mismatch 오류

    [Ubuntu 20.04]RTX 3070, CUDA + cudnn 딥러닝 환경 구성하기

• 환경변수

sudo gedit ~/.bashrc

export CUDA_HOME=/usr/local/cuda-11.3
export PATH=/usr/local/cuda-11.3/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda-11.3/lib64:$LD_LIBRARY_PATH

source ~/.bashrc

• nvcc -V 명령어 입력하면 CUDA 11.3 잘 설치되었는지 확인 가능
  • nvidia-smi에서는 CUDA 버전 다르게 나오는데 무시해도 됨

---

4. cuDNN 8.2.1 설치

Nvidia 로그인해야 다운로드 가능. cuDNN Archive에서 다운로드 받기

tar –xzvf cudnn-11.3-linux-x64-v8.2.1.32.tgz

sudo cp cuda/include/cudnn* /usr/local/cuda-11.3/include
sudo cp cuda/lib64/libcudnn* /usr/local/cuda-11.3/lib64
sudo chmod a+r /usr/local/cuda-11.3/include/cudnn.h /usr/local/cuda-11.3/lib64/libcudnn*

sudo ln -sf /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_adv_train.so.8.2.1 /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_adv_train.so.8
sudo ln -sf /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_ops_infer.so.8.2.1 /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_ops_infer.so.8
sudo ln -sf /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_cnn_train.so.8.2.1 /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_cnn_train.so.8
sudo ln -sf /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_adv_infer.so.8.2.1 /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_adv_infer.so.8
sudo ln -sf /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_ops_train.so.8.2.1 /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_ops_train.so.8
sudo ln -sf /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_cnn_infer.so.8.2.1 /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_cnn_infer.so.8
sudo ln -sf /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn.so.8.2.1 /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn.so.8

sudo ldconfig
ldconfig -N -v $(sed 's/:/ /' <<< $LD_LIBRARY_PATH) 2>/dev/null | grep libcudnn

아래 명령어 입력하면 cudnn 잘 설치되었는지 확인 가능

cat /usr/local/cuda-11.3/include/cudnn_version.h

---

5. OpenPCDet

sudo apt install git

git clone https://github.com/open-mmlab/OpenPCDet.git

conda create -n OpenPCDet python=3.9

conda activate OpenPCDet

pip install torch==1.12.0+cu113 torchvision==0.13.0+cu113 torchaudio==0.12.0 -f https://download.pytorch.org/whl/torch_stable.html

pip install spconv-cu113

pip install -r requirements.txt

python setup.py develop

---

6. OpenPCDet Test

• OpenPCDet 깃허브에서 PV-RCNN 모델 다운로드

  

  OpenPCDet/tools/cfgs/kitti_models/ 안에 넣기
  

• Kitti Dataset 다운로드

  

  하이라이트해둔 4개 다운로드

  OpenPCDet/Data/kitti/ 안에 넣기
  

cd OpenPCDet/tools

python demo.py --cfg_file cfgs/kitti_models/pv_rcnn.yaml --ckpt pv_rcnn_8369.pth --data_path /home/smha/OpenPCDet/data/kitti/training/velodyne/000000.bin