@ Bergen, Norway

第一次安装 CUDA 的过程简直抓狂，中间出现了很多次莫名其妙的 bug，踩了很多坑。比如装好了 CUDA 重启后进不去桌面系统了，直接黑屏、比如鼠标键盘都不 work 了、再比如装好了却安装不了 TensorFlow-GPU......看了一圈网上的安装教程，发现还是官方指南真香了~

新年第一篇，分享一下我的 Ubuntu 18.04 + CUDA 10.0 + cuDNN 7.6.5 + TensorFlow 2.0 安装笔记，希望可以帮助大家少踩坑。

整个安装流程大致是：安装显卡驱动 -> 安装 CUDA^[1] -> 安装 cuDNN^[2] -> 安装 tensorflow-gpu 并测试。

全文目录：

1. Ubuntu安装与更新

2. 安装显卡驱动

3. 安装CUDA

4. 安装cuDNN

5. 安装TensorFlow2.0 GPU及测试

---

1. Ubuntu安装和更新

先进行Ubuntu18.04系统一些基本的安装和更新，具体的操作系统安装过程省略，比较容易，大家可自行百度，有很多教程。

sudo apt-get update # 更新源
sudo apt-get upgrade # 更新已安装的包
sudo apt-get install vim

2. 安装显卡驱动

2.1 禁用 Nouveau 驱动

注意：Linux 系统下有两种方案安装 CUDA：一种是 Package Manager Installation (.deb)，另一种是 Runfile Installation (.run)。本文采取的是第一种（也是官方推荐的方式）。如果使用deb方式安装CUDA可以忽略此步，本人测试OK。如果使用 runfile 安装CUDA需要手动禁用系统自带的 Nouveau 驱动：

lsmod | grep nouveau # 要确保这条命令无输出

vim /etc/modprobe.d/blacklist-nouveau.conf
# 添加下面两行：
#######################################################
blacklist nouveau
options nouveau modeset=0
#######################################################
# 保存后重启：
sudo update-initramfs -u
sudo reboot
# 再次输入以下命令，无输出就表示设置成功了
lsmod | grep nouveau

2.2 安装合适的显卡驱动^[3]

# 先清空现有的显卡驱动及依赖并重启
sudo apt-get remove --purge nvidia*
sudo apt autoremove
sudo reboot

# 添加ppa源并安装最新的驱动
sudo add-apt-repository ppa:graphics-drivers/ppa
sudo apt update
ubuntu-drivers devices
sudo apt install nvidia-driver-440
# 为了防止自动更新驱动导致的兼容性问题，我们还可以锁定驱动版本:
sudo apt-mark hold nvidia-driver-440
# nvidia-driver-440 set on hold.

并在【软件和更新】菜单中的附加驱动列表中，可以找到刚刚安装的nvidia-driver-440，选定即可。输入sudo reboot重启后，输入nvidia-smi，显示下图信息，这样表示显卡驱动已经 ready：

lsmod | grep nvidia # 看到下面的输出则为安装成功，如果无输出，表示有问题

也可以手动去官网下载对应的安装程序安装显卡^[4]

# 动态监测显卡使用的方式：
watch -n 1 nvidia-smi # 1表示每1秒刷新一次
watch -n 0.01 nvidia-smi # 也可改成0.01s刷新一次
# 也可以用gpustat
pip install gpustat
gpustat -i 1 -P

3. 安装 CUDA

百度百科：CUDA（Compute Unified Device Architecture），是显卡厂商NVIDIA^[5]推出的运算平台。CUDA 是一种由 NVIDIA 推出的通用并行计算^[6]架构，该架构使GPU^[7]能够解决复杂的计算问题。

Linux 系统下有两种方案安装 CUDA：一种是 Package Manager Installation (.deb)，另一种是 Runfile Installation (.run)。本文采取的是第一种（也是官方推荐的方式）。

另外，CUDA 对于系统环境有严格的依赖，比如对于 CUDA10.0 有如下的要求。其他的版本可查看对应的Online Documentation^[8]。

3.1 安装前的准备

在安装 CUDA 之前需要先确定环境是 ready 的，以免出现乱七八糟的 bug 无从下手。直接引用官网的说明：

Some actions must be taken before the CUDA Toolkit and Driver can be installed on Linux:

• Verify the system has a CUDA-capable GPU.

• Verify the system is running a supported version of Linux.

• Verify the system has gcc installed.

• Verify the system has the correct kernel headers and development packages installed.

• Download the NVIDIA CUDA Toolkit.

• Handle conflicting installation methods.

3.1.1 确认你有支持 CUDA 的 GPU

lspci | grep -i nvidia | grep VGA

3.1.2 确认你的 linux 版本

uname -m && cat /etc/*release
uname -a
# The x86_64 line indicates you are running on a 64-bit system.

3.1.3 确认 gcc 版本

gcc --version
# gcc (Ubuntu 7.4.0-1ubuntu1~18.04.1) 7.4.0

3.1.4 安装对应内核版本的头文件

查看 kernel 的版本：

uname -r
# 5.0.0-37-generic

This is the version of the kernel headers and development packages that must be installed prior to installing the CUDA Drivers.

安装对应内核版本的头文件：

sudo apt-get install linux-headers-$(uname -r)

3.1.5 选择安装方式

下载对应的安装包（以官方推荐的 Deb packages 安装方式为例）^[9]

The CUDA Toolkit can be installed using either of two different installation mechanisms: distribution-specific packages (RPM and Deb packages), or a distribution-independent package (runfile packages).

(1) The distribution-independent package has the advantage of working across a wider set of Linux distributions, but does not update the distribution's native package management system.

(2) The distribution-specific packages interface with the distribution's native package management system. It is recommended to use the distribution-specific packages, where possible.

3.1.6 彻底卸载之前安装过的相关应用，避免冲突

如果是全新的 ubuntu，可忽略此部分，执行 3.2 部分即可。

如果 ubuntu 下用 RPM/Deb 安装的：

sudo apt-get --purge remove <package_name>
sudo apt autoremove

如果是 runfile 安装的：

sudo /usr/bin/nvidia-uninstall
sudo /usr/local/cuda-X.Y/bin/uninstall_cuda_X.Y.pl

3.2 安装

首先确保已经下载好对应的.deb 文件，然后执行：

sudo dpkg -i cuda-repo-ubuntu1804-10-0-local-10.0.130-410.48_1.0-1_amd64.deb
sudo apt-key add /var/cuda-repo-<version>/7fa2af80.pub # 根据执行完第一步的提示输入，比如我是：
# sudo apt-key add /var/cuda-repo-10-0-local-10.0.130-410.48/7fa2af80.pub
sudo apt-get update
sudo apt-get install cuda-toolkit-10-0 # 注意不是cuda，因为在第二步中装过驱动了，此过程安装cuda-toolkit-10-0即可

3.3 安装后

安装之后需要手动进行一些设置才能使 CUDA 正常的工作。

export PATH=/usr/local/cuda-10.0/bin${PATH:+:${PATH}}

nvcc -V # 检查CUDA是否安装成功
# OUTPUT:
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

最好关闭系统的自动更新，防止安装好的环境突然 bug：

sudo vi /etc/apt/apt.conf.d/10periodic# 修改为：
APT::Periodic::Update-Package-Lists "0";
APT::Periodic::Download-Upgradeable-Packages "0";
APT::Periodic::AutocleanInterval "0";

也可以通过桌面设置：System Settings => Software&Updates => updates

4. 安装 cuDNN^[10]

NVIDIA cuDNN 是用于深度神经网络的 GPU 加速库。首先需要注册下载对应 CUDA 版本号的 cuDNN 安装包: 链接^[11]。

比如对应 CUDA10.0，我下载的是：tar -zxvf cudnn-10.0-linux-x64-v7.6.5.32.tgz

tar -zxvf cudnn-10.0-linux-x64-v7.6.5.32.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*

验证是否安装成功：

cat /usr/local/cuda/include/cudnn.h | grep CUDNN_MAJOR -A 2
# 输出
"""
#define CUDNN_MAJOR 7
#define CUDNN_MINOR 6
#define CUDNN_PATCHLEVEL 5
--
#define CUDNN_VERSION (CUDNN_MAJOR * 1000 + CUDNN_MINOR * 100 + CUDNN_PATCHLEVEL)
#include "driver_types.h"
"""

更推荐使用 Debian File 去安装，因为可以通过里面的样例去验证 cuDNN 是否成功安装。首先下载下面三个文件：

# 分别下载
sudo dpkg -i libcudnn7_7.6.5.32-1+cuda10.0_amd64.deb
sudo dpkg -i libcudnn7-dev_7.6.5.32-1+cuda10.0_amd64.deb
sudo dpkg -i libcudnn7-doc_7.6.5.32-1+cuda10.0_amd64.deb
# 安装完验证：
cp -r /usr/src/cudnn_samples_v7/ $HOME
cd  $HOME/cudnn_samples_v7/mnistCUDNN
make clean && make
./mnistCUDNN
# Test passed!

另外也可以用 conda 来安装 cudatoolkit 和 cuDNN，但要保证驱动是 ready 的。

conda install cudatoolkit=10.0
conda install -c anaconda cudnn

5. 安装 TensorFlow2.0 GPU及测试

# 安装conda
wget https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh && bash Miniconda3-latest-Linux-x86_64.sh
source ~/.bashrc
conda create -y -n tf2 python=3.7
conda activate tf2
pip install --upgrade pip
pip config set global.index-url https://pypi.tuna.tsinghua.edu.cn/simple
pip install tensorflow-gpu
pip install catboost

测试:

import tensorflow as tf
print(tf.__version__)
print("Num GPUs Available: ", len(tf.config.experimental.list_physical_devices('GPU')))
"""
2.0.0
Num GPUs Available:  2
"""

"""
测试程序：
源链接：https://github.com/dragen1860/TensorFlow-2.x-Tutorials/blob/master/08-ResNet/main.py
"""
import os
os.environ["CUDA_VISIBLE_DEVICES"] = "1" # os.environ["CUDA_VISIBLE_DEVICES"] = "0,1"
import tensorflow as tf
import numpy as np
from tensorflow import kerastf.random.set_seed(22)
np.random.seed(22)
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
assert tf.__version__.startswith('2.')(x_train, y_train), (x_test, y_test) = keras.datasets.fashion_mnist.load_data()
x_train, x_test = x_train.astype(np.float32) / 255., x_test.astype(np.float32) / 255.
# [b, 28, 28] => [b, 28, 28, 1]
x_train, x_test = np.expand_dims(x_train, axis=3), np.expand_dims(x_test,axis=3)
# one hot encode the labels. convert back to numpy as we cannot use a combination of numpy
# and tensors as input to keras
y_train_ohe = tf.one_hot(y_train, depth=10).numpy()
y_test_ohe = tf.one_hot(y_test, depth=10).numpy()print(x_train.shape, y_train.shape)
print(x_test.shape, y_test.shape)# 3x3 convolution
def conv3x3(channels, stride=1, kernel=(3, 3)):return keras.layers.Conv2D(channels,kernel,strides=stride,padding='same',use_bias=False,kernel_initializer=tf.random_normal_initializer())class ResnetBlock(keras.Model):def __init__(self, channels, strides=1, residual_path=False):super(ResnetBlock, self).__init__()self.channels = channelsself.strides = stridesself.residual_path = residual_pathself.conv1 = conv3x3(channels, strides)self.bn1 = keras.layers.BatchNormalization()self.conv2 = conv3x3(channels)self.bn2 = keras.layers.BatchNormalization()if residual_path:self.down_conv = conv3x3(channels, strides, kernel=(1, 1))self.down_bn = tf.keras.layers.BatchNormalization()def call(self, inputs, training=None):residual = inputsx = self.bn1(inputs, training=training)x = tf.nn.relu(x)x = self.conv1(x)x = self.bn2(x, training=training)x = tf.nn.relu(x)x = self.conv2(x)# this module can be added into self.# however, module in for can not be added.if self.residual_path:residual = self.down_bn(inputs, training=training)residual = tf.nn.relu(residual)residual = self.down_conv(residual)x = x + residualreturn xclass ResNet(keras.Model):def __init__(self, block_list, num_classes, initial_filters=16, **kwargs):super(ResNet, self).__init__(**kwargs)self.num_blocks = len(block_list)self.block_list = block_listself.in_channels = initial_filtersself.out_channels = initial_filtersself.conv_initial = conv3x3(self.out_channels)self.blocks = keras.models.Sequential(name='dynamic-blocks')# build all the blocksfor block_id in range(len(block_list)):for layer_id in range(block_list[block_id]):if block_id != 0 and layer_id == 0:block = ResnetBlock(self.out_channels,strides=2,residual_path=True)else:if self.in_channels != self.out_channels:residual_path = Trueelse:residual_path = Falseblock = ResnetBlock(self.out_channels,residual_path=residual_path)self.in_channels = self.out_channelsself.blocks.add(block)self.out_channels *= 2self.final_bn = keras.layers.BatchNormalization()self.avg_pool = keras.layers.GlobalAveragePooling2D()self.fc = keras.layers.Dense(num_classes)def call(self, inputs, training=None):out = self.conv_initial(inputs)out = self.blocks(out, training=training)out = self.final_bn(out, training=training)out = tf.nn.relu(out)out = self.avg_pool(out)out = self.fc(out)return outdef main():num_classes = 10batch_size = 128epochs = 2# build model and optimizermodel = ResNet([2, 2, 2], num_classes)model.compile(optimizer=keras.optimizers.Adam(0.001),loss=keras.losses.CategoricalCrossentropy(from_logits=True),metrics=['accuracy'])model.build(input_shape=(None, 28, 28, 1))print("Number of variables in the model :", len(model.variables))model.summary()# trainmodel.fit(x_train,y_train_ohe,batch_size=batch_size,epochs=epochs,validation_data=(x_test, y_test_ohe),verbose=1)# evaluate on test setscores = model.evaluate(x_test, y_test_ohe, batch_size, verbose=1)print("Final test loss and accuracy :", scores)if __name__ == '__main__':main()

监测 GPU 使用：

watch -n 0.01 nvidia-smi

测试 catboost 使用 CPU：

from catboost.datasets import titanic
import numpy as np
from sklearn.model_selection import train_test_split
from catboost import CatBoostClassifier, Pool, cv
from sklearn.metrics import accuracy_scoretrain_df, test_df = titanic()
null_value_stats = train_df.isnull().sum(axis=0)
null_value_stats[null_value_stats != 0]train_df.fillna(-999, inplace=True)
test_df.fillna(-999, inplace=True)X = train_df.drop('Survived', axis=1)
y = train_df.SurvivedX_train, X_validation, y_train, y_validation = train_test_split(X, y, train_size=0.75, random_state=42)
X_test = test_dfcategorical_features_indices = np.where(X.dtypes != np.float)[0]model = CatBoostClassifier(task_type="GPU",custom_metric=['Accuracy'],random_seed=666,logging_level='Silent'
)model.fit(X_train, y_train,cat_features=categorical_features_indices,eval_set=(X_validation, y_validation),logging_level='Verbose',  # you can comment this for no text outputplot=True
);

监测 GPU 使用：

watch -n 0.01 nvidia-smi

REFERENCE

[1]

安装CUDA: https://developer.nvidia.com/cuda-toolkit-archive

[2]

安装cuDNN: https://developer.nvidia.com/rdp/cudnn-download

[3]

安装合适的显卡驱动: http://www.linuxandubuntu.com/home/how-to-install-latest-nvidia-drivers-in-linux

[4]

也可以手动去官网下载对应的安装程序安装显卡: https://www.geforce.cn/drivers

[5]

NVIDIA: https://baike.baidu.com/item/NVIDIA

[6]

并行计算: https://baike.baidu.com/item/并行计算/113443

[7]

GPU: https://baike.baidu.com/item/GPU

[8]

Online Documentation: https://developer.nvidia.com/cuda-toolkit-archive

[9]

下载对应的安装包（以官方推荐的Deb packages安装方式为例）: https://developer.nvidia.com/cuda-10.0-download-archive?target_os=Linux&target_arch=x86_64&target_distro=Ubuntu&target_version=1804&target_type=deblocal

[10]

安装cuDNN: https://developer.nvidia.com/rdp/cudnn-download

[11]

链接: https://developer.nvidia.com/rdp/cudnn-download

[12]

官方-NVIDIA CUDA Installation Guide for Linux: https://docs.nvidia.com/cuda/archive/10.0/cuda-installation-guide-linux/index.html

[13]

CUDA_Quick_Start_Guide-pdf: https://developer.download.nvidia.com/compute/cuda/10.0/Prod/docs/sidebar/CUDA_Quick_Start_Guide.pdf

[14]

CUDA_Installation_Guide_Linux-pdf: https://developer.download.nvidia.com/compute/cuda/10.0/Prod/docs/sidebar/CUDA_Installation_Guide_Linux.pdf

[15]

官方-cuDNN安装: https://docs.nvidia.com/deeplearning/sdk/cudnn-install/index.html#install-linux

[16]

[How To] Install Latest NVIDIA Drivers In Linux: http://www.linuxandubuntu.com/home/how-to-install-latest-nvidia-drivers-in-linux

推荐原创干货阅读：  

聊聊近状， 唠十块钱的

【Deep Learning】详细解读LSTM与GRU单元的各个公式和区别

【手把手AI项目】一、安装win10+linux-Ubuntu16.04的双系统（全网最详细）

【Deep Learning】为什么卷积神经网络中的“卷积”不是卷积运算?

【TOOLS】Pandas如何进行内存优化和数据加速读取（附代码详解）

【TOOLS】python3利用SMTP进行邮件Email自主发送

【手把手AI项目】七、MobileNetSSD通过Ncnn前向推理框架在PC端的使用

【时空序列预测第一篇】什么是时空序列问题？这类问题主要应用了哪些模型？主要应用在哪些领域？

公众号：AI蜗牛车

保持谦逊、保持自律、保持进步

个人微信

备注：昵称+学校/公司+方向

如果没有备注不拉群！

拉你进AI蜗牛车交流群

点个在看,么么哒！