环境信息汇总
- gpu：nVidia GeForce GTX 1050
- Anaconda ：Anaconda 3
- CUDA： 11.0
- cuDNN：cudnn-11.0-windows-x64-v8.0.5.39
- tensorflow_gpu：2.4.0
检查环境信息
【注】 CUDA(ComputeUnified Device Architecture)，是显卡厂商NVIDIA推出的运算平台。 CUDA是一种由NVIDIA推出的通用并行计算架构，该架构使GPU能够解决复杂的计算问题。
是否支持cuda： 根据 forums.developer.nvidia.com/t/cuda-for-… 中回答查看 en.wikipedia.org/wiki/CUDA#G… 有下面描述：
- CUDA SDK 8.0 support for compute capability 2.0 – 6.x (Fermi, Kepler, Maxwell, Pascal). Last version with support for compute capability 2.x (Fermi) (Pascal GTX 1070Ti Not Supported)
- CUDA SDK 9.0 – 9.2 support for compute capability 3.0 – 7.2 (Kepler, Maxwell, Pascal, Volta) (Pascal GTX 1070Ti Not Supported. CUDA SDK 9.0 and support CUDA SDK 9.2).
- CUDA SDK 10.0 – 10.2 support for compute capability 3.0 – 7.5 (Kepler, Maxwell, Pascal, Volta, Turing). Last version with support for compute capability 3.x (Kepler). 10.2 is the last official release for macOS, as support will not be available for macOS in newer releases.
- CUDA SDK 11.0 – 11.2 support for compute capability 3.5 – 8.6 (Kepler (in part), Maxwell, Pascal, Volta, Turing, Ampere)[33] New data types: Bfloat16 and TF32 on third-generations Tensor Cores.[34]**
| Compute capability(version) | GPUs | GeForce | ... |
|---|---|---|---|
| 6.1 | GP102, GP104, GP106, GP107, GP108 | Nvidia TITAN Xp,Titan X,GeForce GTX 1080 Ti, GTX 1080, GTX 1070 Ti, GTX 1070, GTX 1060,GTX 1050 Ti, GTX 1050, GT 1030, GT 1010,MX350, MX330, MX250, MX230, MX150, MX130, MX110 | ... |
安装 Anaconda
Anaconda致力于简化软件包管理系统和部署。Anaconda的包使用软件包管理系统Conda[6]进行管理。超过1200万人使用Anaconda发行版本，并且Anaconda拥有超过1400个适用于Windows、Linux和MacOS的数据科学软件包[7]。
在官网下载安装 www.anaconda.com/ ，并添加环境变量
F:\*\Anaconda3
F:\*\Anaconda3\Scripts
F:\*\Anaconda3\Library\bin
在命令行中执行命令查看是否添加成功:
conda list
CUDA安装
查看 www.tensorflow.org/install/sou…
| 版本 | Python 版本 | 编译器 | 构建工具 | cuDNN | CUDA |
|---|---|---|---|---|---|
| tensorflow_gpu-2.4.0 | 3.6-3.8 | MSVC 2019 | Bazel 3.1.0 | 8.0 | 11.0 |
从此页面进行下载软件包： developer.nvidia.com/cuda-toolki…
测试是否安装成功
C:\Users\xxx>nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2020 NVIDIA Corporation
Built on Thu_Jun_11_22:26:48_Pacific_Daylight_Time_2020
Cuda compilation tools, release 11.0, V11.0.194
Build cuda_11.0_bu.relgpu_drvr445TC445_37.28540450_0
cuDNN 神经网络加速库安装
developer.nvidia.com/zh-cn/cudnn
NVIDIA CUDA® 深度神经网络库 (cuDNN) 是经 GPU 加速的深度神经网络基元库。cuDNN 可大幅优化标准例程（例如用于前向传播和反向传播的卷积层、池化层、归一化层和激活层）的实施。
注册账号登陆后，来到此界面下载与CUDA版本对应的cuDNN： developer.nvidia.com/rdp/cudnn-a…
根据此指导进行复制粘贴
配置环境
可以使用 conda env list 来查看 在 Anaconda 中创建的环境
打开命令行后在windows下可以使用以下命令操作环境：
conda create -n xxx pip python=3.6 创建 python=3.6 的名为xxx环境
activate xxx 来进入 xxx 环境
conda deactivate xxx 退出 xxx 环境
安装 tensorflow
activate xxx 来进入 xxx 环境
使用pip命令来安装对应 tensorflow 版本：
pip install --ignore-installed --upgrade tensorflow_gpu==2.4.0
在xxx环境中运行以下测试代码：
import tensorflow as tf
print("Num GPUs Available: ", len(tf.config.experimental.list_physical_devices('GPU')))
输出为可以使用的gpu个数：
(xxx) C:\Users\30819\Desktop>python 1.py
2021-03-14 17:12:59.896100: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cudart64_110.dll
2021-03-14 17:13:02.240782: I tensorflow/compiler/jit/xla_cpu_device.cc:41] Not creating XLA devices, tf_xla_enable_xla_devices not set
2021-03-14 17:13:02.245444: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library nvcuda.dll
2021-03-14 17:13:03.241515: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1720] Found device 0 with properties:
pciBusID: 0000:01:00.0 name: GeForce GTX 1050 computeCapability: 6.1
coreClock: 1.493GHz coreCount: 5 deviceMemorySize: 4.00GiB deviceMemoryBandwidth: 104.43GiB/s
2021-03-14 17:13:03.251678: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cudart64_110.dll
2021-03-14 17:13:03.262406: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cublas64_11.dll
2021-03-14 17:13:03.266693: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cublasLt64_11.dll
2021-03-14 17:13:03.283413: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cufft64_10.dll
2021-03-14 17:13:03.289394: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library curand64_10.dll
2021-03-14 17:13:03.303645: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cusolver64_10.dll
2021-03-14 17:13:03.310455: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cusparse64_11.dll
2021-03-14 17:13:03.316981: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cudnn64_8.dll
2021-03-14 17:13:03.326594: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1862] Adding visible gpu devices: 0
Num GPUs Available: 1
可以进行更多的尝试代码：
