• TensorFlow安装步骤


    环境:Win10,Python 3.8,Anaconda
    电脑:GeForce RTX 2060, NVIDIA CUDA 11.1.4
    官方教程:在 Windows 上安装 TensorFlow

    1. 下载安装cuda和cudnn

    1.1 下载

    在官网上下载对应的cuda和cudnn,版本可以低于但不能高于电脑支持的版本
    cuda下载地址:cuda-toolkit-archive
    cudnn下载地址:cudnn-archive

    常见问题: 下载cudnn的时候有提示:NVIDIA Developer Program Membership Required
    解决方法: 右键选择【复制链接地址】,把链接复制到迅雷下载器或其他下载软件进行下载
    在这里插入图片描述

    1.2 安装

    CUDA安装:选自定义安装→默认安装路径→安装结束
    安装完成后,会默认生成两个系统变量。
    系统变量查看:此电脑→右键选属性→高级系统设置→环境变量→系统变量
    在这里插入图片描述

    CuDNN安装
    解压→将三个文件夹复制到cuda的安装目录下,直接选择覆盖文件就可以。
    在这里插入图片描述

    安装完了,测试cuda版本。
    打开cmd,输入指令:nvcc -V
    在这里插入图片描述

    2. 创建 tensorflow 环境

    2.1 创建 tensorflow 环境

    conda create -n tensorflow python=3.8
    
    • 1

    常见问题
    ERROR:Collecting package metadata (current——repodata.json): failed
    我的解决方法:将C:\Users\Lenovo下的.condarc文件作如下修改(这里修改一下自己的路径)

    ssl_verify: true
    channels:
      - http://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/win-64/
      - http://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main/win-64
    show_channel_urls: true
    
    • 1
    • 2
    • 3
    • 4
    • 5

    在这里插入图片描述

    2.2 查看环境,可以看到环境已创建**

    conda env list
    
    • 1

    在这里插入图片描述

    2.3 进入环境

    activate tensorflow
    
    • 1

    在这里插入图片描述
    进入具体路径

    # 这里换成自己的具体路径
    >cd C:\Users\Lenovo\anaconda3\envs\tensorflow
    
    • 1
    • 2

    在这里插入图片描述

    2.4 安装默认版本的tensorflow-cpu或者tensorflow-gpu。

    安装tensorflow-cpu版本的,可以输入命令:pip install --ignore-installed --upgrade tensorflow
    安装tensorflow-gpu版本的,可以输入命令:pip install --ignore-installed --upgrade tensorflow-gpu
    在这里插入图片描述
    常见问题
    1 网速问题
    解决方法:直接去网站下载需要的文件(自己电脑下载的是哪个文件就下载哪个文件,我的是tensorflow_gpu-2.9.1-cp38-cp38-win_amd64.whl)
    网址 tensorflow-gpu 2.9.1 ,下载文件放到自己的tensorflow的目录下。
    输入命令:pip install tensorflow_gpu-2.9.1-cp38-cp38-win_amd64.whl
    在这里插入图片描述
    2 ERROR: Could not find a version that satisfies the requirement XXX
    解决方法:直接选用pip源并且信任它的来源就可以解决这种问题

    pip install tensorflow_gpu-2.9.1-cp38-cp38-win_amd64.whl -i http://pypi.douban.com/simple/ --trusted-host pypi.douban.com
    
    • 1

    2.5 查看 tensorflow 版本信息及退出

    查看 tensorflow 版本
    输入命令:pip show tensorflow-gpu,可以查看tensorflow的版本信息
    在这里插入图片描述
    退出 tensorflow 环境

    conda deactivate tensorflow
    
    • 1

    3. 测试 TensorFlow-gpu 是否安装成功

    打开Anaconda,选择tensorflow环境,打开spyder,第一次打开需要安装Spyder,直接点下方的install即可。
    在这里插入图片描述
    测试代码

    import tensorflow as tf
    a = tf.constant(1.)
    b = tf.constant(2.)
    print(a+b)
    print(tf.__version__)
    print(tf.test.gpu_device_name())
    print('GPU:',tf.config.list_physical_devices(device_type='GPU'))
    print('CPU:',tf.config.list_physical_devices(device_type='CPU'))
    print(tf.test.is_gpu_available())
    
    • 1
    • 2
    • 3
    • 4
    • 5
    • 6
    • 7
    • 8
    • 9

    常见问题
    Could not load dynamic library ‘cusolver64_11.dll‘; dlerror: cusolver64_11.dll not found
    解决方法:
    链接:https://pan.baidu.com/s/1W9fR2N_hoVD-7_ODtOiJhg
    提取码:u65i
    下载文件,把文件cusolver64_11.dll添加到创建的环境\Library\bin中
    在这里插入图片描述
    运行代码
    在这里插入图片描述
    安装完成


    3. 测试速度

    import tensorflow as tf
    import timeit
    #指定在cpu上运行
    def cpu_run():
        with tf.device('/cpu:0'):
            cpu_a = tf.random.normal([10000, 1000])
            cpu_b = tf.random.normal([1000, 2000])
            cpu_c = tf.matmul(cpu_a, cpu_b)
            # print( "cpu_a: ", cpu_a.device)
            # print( "cpu_b: ", cpu_b.device)
            # print("cpu_c:", cpu_c.device)
        return cpu_c
    
    #指定在gpu上运行
    
    def gpu_run():
        with tf.device( '/gpu:0'):
            gpu_a = tf.random.normal([ 10000,1000])
            gpu_b = tf.random.normal([ 1000, 2000])
            gpu_c = tf.matmul(gpu_a, gpu_b)
            # print( "gpu_a: ", gpu_a.device)
            # print("gpu_b: ", gpu_b.device)
            # print("gpu_c: ", gpu_c.device)
        return gpu_c
    
    cpu_time = timeit.timeit(cpu_run, number = 10)
    gpu_time = timeit.timeit(gpu_run, number = 10)
    print('cpu:',cpu_time, 'gpu:',gpu_time)
    
    • 1
    • 2
    • 3
    • 4
    • 5
    • 6
    • 7
    • 8
    • 9
    • 10
    • 11
    • 12
    • 13
    • 14
    • 15
    • 16
    • 17
    • 18
    • 19
    • 20
    • 21
    • 22
    • 23
    • 24
    • 25
    • 26
    • 27
    • 28

    在这里插入图片描述
    速度还是快不少的


    参考来源:
    tensorflow详细安装教程(Win10, Anaconda,Python3.9)
    绕过Nvidia官方要求的注册或登陆步骤直接下载cuDNN的方法
    详解Anaconda安装tensorflow报错问题解决方法
    anaconda3创建虚拟环境时出现问题:Collecting package metadata (current——repodata.json): failed

  • 相关阅读:
    File 与 MultipartFile概述
    计算机毕业设计Java东软在线学习平台(源码+系统+mysql数据库+lw文档)
    创建vue3项目并引用elementui
    MODBUS协议下,能否实现MCGS触摸屏与FX5U之间无线通讯?
    “之江创客”聚焦农村电商创新发展 扎实助推共同富裕
    Google Earth Engine ——neighborhoodToBands函数的使用
    E047-论坛漏洞分析及利用-针对Wordpress论坛进行信息收集与漏洞扫描的探索
    多测师肖sir_高级金牌讲师_python之作业006
    【飞桨PaddleSpeech语音技术课程】— 声纹检索系统与实践
    JAVA毕业设计课设源码分享50+例
  • 原文地址:https://blog.csdn.net/weixin_46599926/article/details/126112748