一行代码安装Tensorflow GPU Ubuntu20.04 Windows

本文内容：Ubuntu20.04下使用Anaconda 配置Tensorflow GPU环境

Windows操作系统也可行，换成Win下的anaconda安装方法即可。

Why conda？

使用conda安装Tensorflow GPU应该是最简单的安装方法，conda会自动匹配tensoflow对应的cuda等环境的版本并安装。

安装Anaconda

本人安装的是2022.5版，其他版本可自行在此寻找

wget https://repo.anaconda.com/archive/Anaconda3-2022.05-Linux-x86_64.sh

bash Anaconda3-2022.05-Linux-x86_64.sh 

提示需要阅读许可条例，按下Enter键继续，如图：

出现提示，是否接受许可条例，输入“yes” 按Enter键，在界面中会提示Anaconda将要安装的位置，按下Enter键确认，如图：

提示是否要写入配置文件，输入“yes”按Enter键，如图：

到这里，Anaconda就安装成功了。执行“source ~/.bashrc” 让配置生效，重启终端

检验conda，会输出conda版本

conda --version

更新到最新版本

conda update conda
conda --version
conda 4.13.0

安装显卡驱动

一般都有自带的驱动，这步略，需要的自行安装

创建conda虚拟环境

conda create --name tf_gpu python=3.8
#
# To activate this environment, use
#
#     $ conda activate tf_gpu
#
# To deactivate an active environment, use
#
#     $ conda deactivate

进入环境

conda activate tf_gpu

安装tensorflow-gpu

conda install tensorflow-gpu

此步会自动找到所有所需软件，如图

直接输入 y 享受一键安装

验证

import tensorflow as tf
print(tf.config.list_physical_devices('GPU'))

(tf_gpu) $ python3 test.py 
2022-06-29 05:06:46.545267: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.10.1
2022-06-29 05:07:01.046758: I tensorflow/compiler/jit/xla_cpu_device.cc:41] Not creating XLA devices, tf_xla_enable_xla_devices not set
2022-06-29 05:07:01.047986: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcuda.so.1
2022-06-29 05:07:01.098053: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1720] Found device 0 with properties: 
pciBusID: 0000:18:00.0 name: NVIDIA GeForce RTX 3090 computeCapability: 8.6
coreClock: 1.86GHz coreCount: 82 deviceMemorySize: 23.70GiB deviceMemoryBandwidth: 871.81GiB/s
2022-06-29 05:07:01.098098: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.10.1
2022-06-29 05:07:01.678401: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.10
2022-06-29 05:07:01.678585: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublasLt.so.10
2022-06-29 05:07:02.019572: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2022-06-29 05:07:02.084377: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2022-06-29 05:07:02.692309: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.10
2022-06-29 05:07:02.765532: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.10
2022-06-29 05:07:03.966511: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.7
2022-06-29 05:07:03.968668: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1862] Adding visible gpu devices: 0
[PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU')]

环境成功！