Python Server Build Log

建置環境

1.更新python版本

sudo add-apt-repository ppa:deadsnakes/ppa
sudo apt-get update
sudo apt-get install python3.10
sudo update-alternatives --install /usr/bin/python3 python3 /usr/bin/python3.9 1
sudo update-alternatives --install /usr/bin/python3 python3 /usr/bin/python3.10 2
sudo update-alternatives --config python3

2.清空資料

cd /usr/share/keyrings/
sudo rm -rf /usr/lib/nvidia/
sudo apt-get --purge remove "*cublas*" "*cuda*" "nsight*" "*nvidia*"
sudo apt autoremove
sudo apt autoclean
sudo apt-get install build-essential linux-headers-$(uname -r)

2.安裝nvidia driver

sudo apt-get update
#https://www.nvidia.com/Download/index.aspx?lang=en-us
sudo bash NVIDIA-Linux-x86_64-550.67.run
# sudo apt install nvidia-detect
# sudo apt install nvidia-driver

3.安裝cuda

#https://developer.nvidia.com/cuda-downloads
wget https://developer.download.nvidia.com/compute/cuda/11.7.1/local_installers/cuda-repo-debian11-11-7-local_11.7.1-515.65.01-1_amd64.deb
sudo dpkg -i cuda-repo-debian11-11-7-local_11.7.1-515.65.01-1_amd64.deb
sudo cp /var/cuda-repo-debian11-11-7-local/cuda-*-keyring.gpg /usr/share/keyrings/
sudo add-apt-repository contrib
sudo apt-get update
sudo apt-get -y install cuda
nvidia-smi

4.安裝cudnn

#https://developer.nvidia.com/rdp/cudnn-download
sudo tar -xzvf cudnn-11.3-linux-x64-v8.2.1.32.tgz
sudo cp cuda/include/cudnn.h /usr/local/cuda/include
sudo cp cuda/lib64/libcudnn* /usr/local/cuda/lib64
sudo chmod a+r /usr/local/cuda/include/cudnn.h /usr/local/cuda/lib64/libcudnn*

5.安裝miniconda

$wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
$bash
conda create -n myDL python=3.10 pip
conda activate myDL

6.安裝PyTorch

conda install pytorch torchvision torchaudio cudatoolkit=11.7 -c pytorch -c conda-forge

7.其他指令

pip3 install -U d2l jupyter
sudo apt install nvidia-driver firmware-misc-nonfree
檢驗版本 dpkg -l | grep -i nvidia

8.遠端Jupyter

https://stackoverflow.com/questions/42848130/why-i-cant-access-remote-jupyter-notebook-server
jupyter notebook --no-browser & disown

10.故障排除

perl 語言錯誤
sudo apt-get update && sudo apt-get install locales
sudo dpkg-reconfigure locales

header搜尋
ls /usr/src/linux-6.1.0-headers
sudo apt list linux-headers-$(uname -r)
sudo apt install linux-headers-amd64
#sudo apt install kernel-headers-$(uname -r)

sudo apt install build-essential
sudo dpkg --configure -a
sudo dkms install -m nvidia-current -v 535.104.05 
sudo apt-get install multipath-tools

11.Stable Diffusion

mkdir StableDiff
cd StableDiff

# 程式框架部分
git clone https://github.com/AUTOMATIC1111/stable-diffusion-webui.git

# 下載權重檔
wget https://huggingface.co/runwayml/stable-diffusion-v1-5/resolve/main/v1-5-pruned-emaonly.ckpt
mv v1-5-pruned-emaonly.ckpt stable-diffusion-webui/models/Stable-diffusion/

# 執行
./webui.sh --xformers --share

參考資料

https://docs.conda.io/en/latest/miniconda.html
https://developer.nvidia.com/cuda-downloads
https://developer.nvidia.com/rdp/cudnn-download
https://pytorch.org/get-started/locally/
https://docs.nvidia.com/deploy/cuda-compatibility/index.html
https://www.datacamp.com/tutorial/how-to-run-stable-diffusion