TensorFlow CUDA CuDNN 与 python版本对应关系及参考链接

1.安装流程

1.1 查看自己的显卡类型及是否支持CUDA

查看是否支持CUDA
https://developer.nvidia.com/cuda-gpus

1.2 查看显卡对应驱动

https://www.nvidia.cn/geforce/drivers/

1.3 查看驱动对应CUDA版本

https://docs.nvidia.com/cuda/cuda-toolkit-release-notes/index.html

1.4 查看对应CuDNN及Tensorflow

1.5 下载安装CUDA，CUDNN，Tensorflow

---

安装CUDA，写入环境变量

---

复制CUDNN的include，bin，lib到CUDA对应文件

---

安装Anaconda，安装tensorflow

---

2.参考教程

1.在Ubuntu 18.04上安装Conda TensorFlow-gpu和Keras
2.在Ubuntu 18.04 / 19.04上安装Tensorflow-GPU以在Anaconda上使用Nvidia GPU做AI！
3.Tensorflow GPU的安装变得容易：Ubuntu版本
4.安装Anaconda的TensorFlow-GPU（conda install tensorflow-gpu）
5.Tensorflow2-GPU（Nvidia）+ Ubuntu 18.04 + Anaconda + Jupyter –设置指南
6.Ubuntu Anaconda安装TensorFlow-GPU

3.资料链接

查看NVIDIA支持CUDA功能的GPU
https://developer.nvidia.com/cuda-gpus
Conda 说明书
https://docs.conda.io/en/latest/miniconda.html
Anaconda说明书
https://docs.anaconda.com/
Nvidia 参考文档
https://docs.nvidia.com/cuda/cuda-installation-guide-linux/#axzz4VZnqTJ2A
CUDA 下载
https://developer.nvidia.com/cuda-downloads?target_os=Linux&target_arch=x86_64&target_distro=Ubuntu&target_version=1604
CuDNN下载
https://developer.nvidia.com/rdp/cudnn-download

CUDA与NVIDA驱动对应关系
https://docs.nvidia.com/deploy/cuda-compatibility/index.html#binary-compatibility__table-toolkit-driver
TensorFlow2
https://www.tensorflow.org/install
Tensorflow版本对应
https://www.tensorflow.org/install/source
NVIDIA驱动下载
https://www.nvidia.cn/geforce/drivers/