Pytorch、theano+cuda配置

Pytorch与theano的cuda 配置:

一、pytorch 无需安装cuda与cudnn，直接使用官网命令“conda install pytorch torchvision cudatoolkit==10.2 -c pytorch”(版本自行选择)即可

二、theano cuda加速需要先安装cuda与cudnn（9.0已成功，切记使用.run文件安装，跳过nvidia显卡驱动安装！！），完成安装后再安装theano（conda install theano，直接安装最新版本1.0.5），并且在pycharm的terminal或吃哦那大虚拟环境中输入命令“conda install cudatoolkit 自动安装8.0版本”。最后进行GPU配置，输入命令“sudo gedit ～/.theanorc”,编辑内容如下：
[cuda]
root=/usr/local/cuda-9.0
[global]
device=cuda
floatX=float32
[nvcc]
fastmath=True
完成后采用测试代码：

from theano import function, config, shared, tensor
import numpy
import time
 
vlen = 10 * 30 * 768  # 10 x #cores x # threads per core
iters = 1000
 
rng = numpy.random.RandomState(22)
x = shared(numpy.asarray(rng.rand(vlen), config.floatX))
f = function([], tensor.exp(x))
print(f.maker.fgraph.toposort())
t0 = time.time()
for i in range(iters):
    r = f()
t1 = time.time()
print("Looping %d times took %f seconds" % (iters, t1 - t0))
print("Result is %s" % (numpy.asarray(r),))
if numpy.any([isinstance(x.op, tensor.Elemwise) and
              ('Gpu' not in type(x.op).__name__)
              for x in f.maker.fgraph.toposort()]):
    print('Used the cpu')
else:
    print('Used the gpu')