Win10 Theano Install Guide

最新推荐文章于 2020-12-11 12:38:50 发布

weixin_30439031

最新推荐文章于 2020-12-11 12:38:50 发布

阅读量56

点赞数

CC 4.0 BY-SA版权

原文链接：http://www.cnblogs.com/ZJUT-jiangnan/p/4135059.html

本文详细介绍了如何使用Miniconda安装Theano，并配置其使用GPU加速的方法。包括安装必要的依赖库，如CUDA和Visual Studio 2015等步骤。此外，还提供了验证GPU是否正确配置的示例代码。

basic install guide

1. download miniconda

2. conda install libpython mingw

3. conda install theano

now you can run theano correctly.

Enable GPU

1. download and install cuda8.0

2. downlaod and install vs2015

2. download and install Windows Kits

3. add path "LIB" = "C:\Program Files (x86)\Windows Kits\10\Lib\10.0.14393.0\um\x64;C:\Program Files (x86)\Windows Kits\10\Lib\10.0.14393.0\ucrt\x64"

add path "INCLUDE" = "C:\Program Files (x86)\Windows Kits\10\Include\10.0.14393.0\ucrt";

4. add "C:/user/your_user_name/.theanorc.txt":

1 [global] 
2 profile_optimizer=True
3 profile=True
4 floatX = float32 
5 device = gpu 
6 [nvcc] 
7 compiler_bindir=D:\Program Files (x86)\Microsoft Visual Studio 14.0\VC\bin
8 [dnn]
9 enable=auto

5. theano gpu cpu test:

 1 from theano import function,config,shared,sandbox
 2 import theano.tensor as T 
 3 import numpy 
 4 import time
 5 print config.profile
 6 
 7 vlen=10*30*768
 8 iters=1000
 9 rng=numpy.random.RandomState(22)
10 x=shared(numpy.asarray(rng.rand(vlen),config.floatX))
11 f=function([],T.exp(x))
12 print (f.maker.fgraph.toposort())
13 t0=time.time()
14 for i in range(iters):
15     r=f()
16 t1=time.time()
17 print("loop %d time took "% iters, t1-t0,'seconds')
18 print("results is",r)
19 if numpy.any([isinstance(x.op,T.Elemwise) for x in f.maker.fgraph.toposort()]):
20     print ('used the cpu')
21 else:
22     print ('used the gpu')