Ubuntu18.04安裝tensorflow-gpu(不使用Docker)
版本設(shè)置
tensorflow-gpu:1.14.0、nvidia-driver-418婿脸、cuda-10.0
注意：版本搭配，否則會(huì)導(dǎo)致各種問(wèn)題崭篡。參考tensorflow官網(wǎng)的版本搭配
首先安裝nvidia-driver-418
查看當(dāng)前顯卡驅(qū)動(dòng)信息
lshw -C display | configuration
將nvidia-driver-418 repository添加到apt
#下載cuda deb文件
wget https://developer.download.nvidia.cn/compute/cuda/repos/ubuntu1804/x86_64/cuda-10-0_10.0.130-1_amd64.deb
#根據(jù)deb文件構(gòu)建軟件包
sudo dpkg -i cuda-10-0_10.0.130-1_amd64.deb
#獲取公鑰
sudo apt-key adv --fetch-keys https://developer.download.nvidia.cn/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub
sudo apt update
wget https://developer.download.nvidia.cn/compute/machine-learning/repos/ubuntu1804/x86_64/nvidia-machine-learning-repo-ubuntu1804_1.0.0-1_amd64.deb
sudo apt install ./nvidia-machine-learning-repo-ubuntu1804_1.0.0-1_amd64.deb
sudo apt update
開(kāi)始安裝驅(qū)動(dòng)
#查看上一步在apt中內(nèi)建的nvidia driver，注意版本是否為我們需要安裝的版本號(hào)
ubuntu-drivers devices
#輸出為
== /sys/devices/pci0000:00/0000:00:02.0/0000:03:00.0 ==
modalias : pci:v000010DEd00001B84sv00007377sd00000000bc03sc00i00
vendor : NVIDIA Corporation
model : GP104 [GeForce GTX 1060 3GB]
driver : nvidia-driver-410 - third-party free
driver : nvidia-driver-418 - third-party free recommended
driver : nvidia-driver-390 - distro non-free
driver : xserver-xorg-video-nouveau - distro free builtin
#開(kāi)始安裝
sudo ubuntu-drivers autoinstall
#安裝完成后重啟
sudo reboot
#查看驅(qū)動(dòng)信息
nvidia-smi
#輸出信息為
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 418.67 Driver Version: 418.67 CUDA Version: 10.1 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 GeForce GTX 106... On | 00000000:03:00.0 On | N/A |
| 36% 37C P8 7W / 120W | 434MiB / 3016MiB | 0% Default |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 1072 G /usr/lib/xorg/Xorg 16MiB |
| 0 1138 G /usr/bin/gnome-shell 49MiB |
| 0 1429 G /usr/lib/xorg/Xorg 122MiB |
| 0 1561 G /usr/bin/gnome-shell 155MiB |
| 0 2536 C python3 59MiB |
| 0 3421 G ...quest-channel-token=1415105501360332168 25MiB |
+-----------------------------------------------------------------------------+
驅(qū)動(dòng)安裝后安裝cuda-10.0
下載cuda runfile文件
從官網(wǎng)https://developer.nvidia.com/cuda-10.0-download-archive下載runfile 文件涡上，如圖
2019-07-23 14-38-52屏幕截圖.png
安裝cuda
下載完成后，運(yùn)行文件
sudo sh cuda_10.0.130.410.48_linux.run
根據(jù)提示進(jìn)行安裝拒名，跳過(guò)驅(qū)動(dòng)安裝部分吩愧。
安裝成功后會(huì)生成/usr/local/cuda-10.0文件夾
添加環(huán)境變量
sudo vim /etc/profile
#添加下面兩條語(yǔ)句到文件中
export PATH=/usr/local/cuda-10.0/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda-10.0/lib64:$LD_LIBRARY_PATH
#重啟生效
sudo reboot
#查看cuda 版本
nvcc --version
#輸出結(jié)果為
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2018 NVIDIA Corporation
Built on Sat_Aug_25_21:08:01_CDT_2018
Cuda compilation tools, release 10.0, V10.0.130
#至此cuda10.0安裝成功
安裝cudnn
從官網(wǎng)上下載最新版本的cudnn：https://developer.nvidia.com/rdp/cudnn-archive
注意版本搭配
下載后，進(jìn)行壓縮包放置的文件夾
tar xvzf cudnn-10.1-linux-x64-v7.6.1.34.tgz
sudo cp /cuda/include/* /usr/local/cuda-10.0/include/
sudo cp /cuda/lib64/* /usr/local/cuda-10.0/lib64/
sudo chmod a+r /usr/local/cuda-10.0/include/cudnn.h /usr/local/cuda-10.0/lib64/libcudnn*
安裝tensorflow
#安裝最穩(wěn)定版本的tensorflow-gpu,版本號(hào)為1.14.0
pip3 install tensorflow-gpu
測(cè)試tensorflow
運(yùn)行任意一個(gè)使用到tensoflow的文件增显，輸出結(jié)果正確則測(cè)試通過(guò)雁佳。
