> 安裝ubuntu 18.04要記得選取"安裝第三方驅動" # Ubuntu 18.04 搭建深度學習開發環境_CUDA 11.0_cuDNN 8.0.2_TensorRT 7.1.3_TensorFlow安裝教學 ## Step1.　440顯卡驅動(灌完cuda 11.0 後會自動改成450驅動) 1.重灌完後，在grub界面按e進入修改參數，quite splash 後面空一格後加入nomodeset後進入系統 2.利用ubuntu自帶的"軟體與更新"，選擇欲用的"顯卡驅動" 3.之後也不需要永久修改grub參數了，重啟後仍然會使用"專用顯卡驅動" ## Step2.　藍芽、聲音控制界面 1.bluez從5.48更新到5.50 ``` dpkg --status bluez | grep '^Version:'` #查看bluez版本 sudo add-apt-repository ppa:bluetooth/bluez #添加套件源 sudo apt-get update sudo apt upgrade #進入藍芽界面scan on、pair和trust mac地址 bluetoothctl scan on pair mac地址 trust mac 地址 ``` 2.聲音控制界面(開啟界面後將"線路輸入"改成"耳機"就可以讓後方面板音源線有聲音) ``` sudo apt install pavucontrol pavucontrol #開啟聲音控制界面（因為gnome內建的再後方面板音源線預設設定有Bug） ``` ## Step3.　安裝pip 3、conda 1.pip 3 ``` sudo apt install python3-pip pip3 install setuptools ``` 2.conda [下載](https://www.anaconda.com/products/individual) [參考源](https://docs.conda.io/projects/conda/en/latest/user-guide/install/linux.html) ``` bash Anaconda3-2020.07-Linux-x86_64.sh ``` ## Step4.　CUDA 11.0在turing顯卡下apt安裝 1.安裝前硬體資訊檢查 [[參考源]](https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#pre-installation-actions) 2.進行Runfile Installation [[參考源]](https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#pre-installation-actions) ``` lsmod | grep nouveau #確保沒有任何文字被輸出（詳細請見參考源） ``` >這邊不繼續下去～因為比較喜歡用套件管理軟體安裝 3.進行package Manager Installation [[參考源]](https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#pre-installation-actions) ``` wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-ubuntu1804.pin sudo mv cuda-ubuntu1804.pin /etc/apt/preferences.d/cuda-repository-pin-600 wget http://developer.download.nvidia.com/compute/cuda/11.0.2/local_installers/cuda-repo-ubuntu1804-11-0-local_11.0.2-450.51.05-1_amd64.deb sudo dpkg -i cuda-repo-ubuntu1804-11-0-local_11.0.2-450.51.05-1_amd64.deb sudo apt-key add /var/cuda-repo-ubuntu1804-11-0-local/7fa2af80.pub sudo apt-get update sudo apt-get install --no-install-recommends cuda ``` 4.新增環境變數至檔案最後面 [[參考源]](https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#post-installation-actions) ``` vim ~/.bashrc ``` * export PATH=/usr/local/cuda-11.0/bin${PATH:+:${PATH}} * ~~export LDLIBRARYPATH="/usr/local/cuda-11.0/lib64:${LDLIBRARYPATH}"~~ ``` source ~/.bashrc #載入新的bash設定檔 ``` 5.驗證安裝是否成功 Verify the Driver Version ``` cat /proc/driver/nvidia/version #出現kernel version表成功 ``` 5.1 查看NVIDIA-SMI和Driver Version的版本號是否一致 ``` nvidia-smi #順便看CUDA Version是不是你裝好的！ ``` 應該會顯示error，因為cuda11自動裝了相容的顯卡驅動450，須重新開機讓系統應用。 5.2 重開機再確認一次版本號 ``` reboot nvidia-smi #看看NVIDIA-SMI和Driver Version是否一致 ``` 應該一致了，如果不一致，那肯定沒做好之前的步驟。 6.（選擇性安裝套件 #我有裝，但好像沒必要） ``` sudo apt-get install g++ freeglut3-dev build-essential libx11-dev \ libxmu-dev libxi-dev libglu1-mesa libglu1-mesa-dev ``` 7.(當你想要移除cuda) ``` sudo apt-get --purge remove "*cublas*" "*cufft*" "*curand*" \ "*cusolver*" "*cusparse*" "*npp*" "*nvjpeg*" "cuda*" "nsight*" ``` 8.(當你想要移除NVIDIA driver) ``` sudo apt-get --purge remove "*nvidia*" ``` ## Step5.　cuDNN 8.0.2在cuda 11.0下(.deb)安裝 0.下載cuDNN v8.0.2 (July 24th, 2020)forCUDA 11.0 [註冊後下載](https://developer.nvidia.com/rdp/cudnn-download) |cuDNN Runtime Library for Ubuntu18.04 x86_64 (Deb) |cuDNN Developer Library for Ubuntu18.04 x86_64 (Deb) |cuDNN Code Samples and User Guide for Ubuntu18.04 x86_64 (Deb) 1.安裝cudnn庫 ``` sudo dpkg -i libcudnn8_8.0.2.39-1+cuda11.0_amd64.deb ``` 2.安裝開發者cudnn庫 ``` sudo dpkg -i libcudnn8-dev_8.0.2.39-1+cuda11.0_amd64.deb ``` 3.驗證cudnn和cuda成功運行 ``` sudo dpkg -i libcudnn8-doc_8.0.2.39-1+cuda11.0_amd64.deb cp -r /usr/src/cudnn_samples_v8/ $HOME cd $HOME/cudnn_samples_v8/mnistCUDNN make clean && make #會出現很多warning沒差拉！ ./mnistCUDNN # "Test passed!" 出現表示成功 ``` ## Step6.　TensorRT 7.1.3.4在cuda 11.0下(.deb)安裝 0.下載TensorRT 7.1 GA [註冊後下載](https://developer.nvidia.com/nvidia-tensorrt-7x-download) |TensorRT 7.1.3.4 for Ubuntu 1804 and CUDA 11.0 DEB local repo packages 1.確認RT支援的推理精度和特殊硬體的支援功能[參考源](https://docs.nvidia.com/deeplearning/tensorrt/archives/tensorrt-713/support-matrix/index.html) 2.先安裝PyCUDA[參考源](https://docs.nvidia.com/deeplearning/tensorrt/archives/tensorrt-713/install-guide/index.html#installing-pycuda) ``` nvcc --version #確認nvcc能被bash找到 ``` 顯示以下資訊表示成功 ```ShellSession nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2020 NVIDIA Corporation Built on Thu_Jun_11_22:26:38_PDT_2020 Cuda compilation tools, release 11.0, V11.0.194 Build cuda_11.0_bu.TC445_37.28540450_0 ``` ``` pip3 install 'pycuda>=2019.1.1' ``` 3.開始TensorRT(.deb)安裝 [參考源](https://docs.nvidia.com/deeplearning/tensorrt/archives/tensorrt-713/install-guide/index.html#installing-debian) ``` cd 至"nv-tensorrt-repo-ubuntu1804-cuda11.0-trt7.1.3.4-ga-20200617_1-1_amd64.deb" 檔案資料夾位置 sudo dpkg -i nv-tensorrt-repo-ubuntu1804-cuda11.0-trt7.1.3.4-ga-20200617_1-1_amd64.deb sudo apt-key add /var/nv-tensorrt-repo-cuda11.0-trt7.1.3.4-ga-20200617/7fa2af80.pub sudo apt-get update sudo apt-get install tensorrt cuda-nvrtc-11-0 sudo apt-get install python3-libnvinfer-dev sudo apt-get install uff-converter-tf ``` 4.驗證TensorRT安裝 ``` dpkg -l | grep TensorRT ``` 顯示以下資訊表示成功 ```ShellSession ii graphsurgeon-tf 7.1.3-1+cuda11.0 amd64 GraphSurgeon for TensorRT package ii libnvinfer-bin 7.1.3-1+cuda11.0 amd64 TensorRT binaries ii libnvinfer-dev 7.1.3-1+cuda11.0 amd64 TensorRT development libraries and headers ii libnvinfer-doc 7.1.3-1+cuda11.0 all TensorRT documentation ii libnvinfer-plugin-dev 7.1.3-1+cuda11.0 amd64 TensorRT plugin libraries ii libnvinfer-plugin7 7.1.3-1+cuda11.0 amd64 TensorRT plugin libraries ii libnvinfer-samples 7.1.3-1+cuda11.0 all TensorRT samples ii libnvinfer7 7.1.3-1+cuda11.0 amd64 TensorRT runtime libraries ii libnvonnxparsers-dev 7.1.3-1+cuda11.0 amd64 TensorRT ONNX libraries ii libnvonnxparsers7 7.1.3-1+cuda11.0 amd64 TensorRT ONNX libraries ii libnvparsers-dev 7.1.3-1+cuda11.0 amd64 TensorRT parsers libraries ii libnvparsers7 7.1.3-1+cuda11.0 amd64 TensorRT parsers libraries ii python3-libnvinfer 7.1.3-1+cuda11.0 amd64 Python 3 bindings for TensorRT ii python3-libnvinfer-dev 7.1.3-1+cuda11.0 amd64 Python 3 development package for TensorRT ii tensorrt 7.1.3.4-1+cuda11.0 amd64 Meta package of TensorRT ii uff-converter-tf 7.1.3-1+cuda11.0 amd64 UFF converter for TensorRT package ``` 5.如果你要當App Server用來推理 ``` sudo apt-get update sudo apt-get install libnvinfer7 cuda-nvrtc-11-0 sudo apt-get install python3-libnvinfer ``` ## Step7.　TensorFlow 安裝 安裝前重啟一次電腦 [參考源](https://www.tensorflow.org/install/pip) ``` sudo apt update sudo apt install python3-dev python3-pip sudo pip3 install -U virtualenv # system-wide install #建立虛擬環境 virtualenv --system-site-packages -p python3 ./venv source ./venv/bin/activate # sh, bash, ksh, or zsh pip install --upgrade pip pip install --upgrade tensorflow #驗證tensorflow安裝 python -c "import tensorflow as tf;print(tf.reduce_sum(tf.random.normal([1000, 1000])))" #要關掉虛擬環境（注意tensorflow不能在運行的情況下關閉） deactivate # don't exit until you're done using TensorFlow ``` ## 我們成功裝好CUDA；cuDNN；TensorRT;TensorFlow囉！ ## Step8.　SSH安裝 1.安裝 ``` sudo apt update sudo apt install openssh-server ``` 2.驗證 ``` sudo systemctl status ssh ``` 顯示包含以下資訊表示成功 ```ShellSession Active: active (running) ``` 按下"q"離開 3.打開ssh port ``` sudo ufw allow ssh ``` 4.如果想關閉ssh ``` sudo systemctl stop ssh sudo systemctl disable ssh ```