warp-CTC安装
http://www.pianshen.com/article/862428080/
warpctc_pytorch 编译不成功的解决办法
warp-CTC是百度开源的一个可以应用在CPU和GPU上高效并行的CTC代码库，对CTC算法进行了并行处理。
warp-CTC安装：
git clone https://github.com/SeanNaren/warp-ctc.git
cd warp-ctc
mkdir build; cd build
cmake ..
make
cd ../pytorch_binding
python setup.py install
添加环境变量：
gedit ./.bashrc
export WARP_CTC_PATH=/home/xxx/warp-ctc/build
验证pytorch中warp-CTC是否可用GPU例子：
cd /home/xxx/warp-ctc/pytorch_binding/tests
python test_gpu.py
OK输出：
或：
import torch
from torch.autograd import Variable
from warpctc_pytorch import CTCLoss
ctc_loss = CTCLoss()
# expected shape of seqLength x batchSize x alphabet_size
probs = torch.FloatTensor([[[0.1, 0.6, 0.1, 0.1, 0.1], [0.1, 0.1, 0.6, 0.1, 0.1]]]).transpose(0, 1).contiguous()
labels = Variable(torch.IntTensor([1, 2]))
label_sizes = Variable(torch.IntTensor([2]))
probs_sizes = Variable(torch.IntTensor([2]))
probs = Variable(probs, requires_grad=True) # tells autograd to compute gradients for probs
cost = ctc_loss(probs, labels, probs_sizes, label_sizes)
cost.backward()
print('PyTorch bindings for Warp-ctc')
PyTorch bindings for Warp-ctc参考：https://github.com/SeanNaren/warp-ctc
---------------------
原文：https://blog.csdn.net/dcrmg/article/details/80199722
========win10 +pytorch
在CMakeLists.txt中添加此行：
set（CUDA_curand_LIBRARY“C：/ Program Files / NVIDIA GPU Computing Toolkit / CUDA / v9.2 / lib / x64 / curand.lib”）
如果发生'Dll Load Failed'错误，请复制生成的warpctc.dll并从torch目录复制caffe2_gpu.dll，caffe2.dll，粘贴到安装目录。
与VS 2017一起构建
git clone https://github.com/amberblade/warp-ctc
cd warp-ctc
mkdir build;
cd build
cmake -G "Visual Studio 15 2017 Win64" ..现在去构建目录，打开sln文件并构建解决方案
然后安装绑定：
cd pytorch_binding
python setup.py install如果您尝试上述并获得“DLL加载失败”错误：
cd ../pytorch_binding
python setup.py install
cd ../build
cp warpctc.dll /path/to/your/warp-ctc/installed
# find your caffe2_gpu.dll and caffe2.dll, and copy to /path/to/your/warp-ctc/installedWrap-CTC的简单编译（只涉及cpu部分 linux） https://blog.csdn.net/qq_26819733/article/details/53608308