These days, not a bit of artificial intelligence and neural network, are embarrassed to greet people. Before doing a bit of sklearn, today think should understand Google this legendary AI open source artifact.
Finally there is time, everything from Hello World, start with the environment.
1. Python Environment
Installing python:https://www.python.org/
Install a variety of libraries, depending on, because already loaded too much, I have forgotten what needs to be loaded. But sometimes the wheel will come in handy if it doesn't fit.
Various wheels: http://www.lfd.uci.edu/~gohlke/pythonlibs/
Wheel Mount Command:
Pip Install \PATH\TO\DOWNLOADED\FILE\*.WHL
2. Install Visual Studio
Yes, to install Visual Studio first, this to the official website to download, loaded on the line.
Official website: https://www.visualstudio.com/
3. Installation of Cuda and CUDNN
Cuda:https://developer.nvidia.com/cuda-downloads
This is the right installation.
Cudnn:https://developer.nvidia.com/rdp/cudnn-download
After this decompression, the bin path is added to the path inside to get it done.
Verify that the installation is successful:
Nvcc-v
4, Installation TensorFlow
PIP3 Install--upgrade TensorFlow
PIP3 Install--upgrade Tensorflow-gpu
Verify that the installation is successful, go straight to the Python command line interface, import TensorFlow, no error, that is, the installation is successful.
Import TensorFlow
5, finish the call.
6. Reference
Http://www.tensorfly.cn/tfdoc/get_started/os_setup.html
Https://www.tensorflow.org/install/install_windows
http://blog.csdn.net/u010099080/article/details/53418159
http://blog.csdn.net/include1224/article/details/53452824
Setting up a TensorFlow environment under windows