之前LZ在编译tf1.14接口的时候,其实就已经强调了Eigen 版本选择的重要性,但是我们经常使用不止一种三方库,在使用不同算法库的时候其实就会存在这样那样动态库冲突的情况,而且这种情况最为头疼。。。
下面是LZ的一段测试代码
#include <stdio.h>
#include <iostream>
...
tf的一些头文件
...
#include "opencv2/opencv.hpp"
#include "opencv2/core/cuda.hpp"
#include "opencv2/cudawarping.hpp"
#include "opencv2/gpu/gpu.hpp"
int main() {
int num_devices = cv::cuda::getCudaEnabledDeviceCount();
if (num_devices <= 0) {
std::cerr << "There is no device." << std::endl;
return -1;
}
int enable_device_id = -1;
for (int i = 0; i < num_devices; i++) {
cv::cuda::DeviceInfo dev_info(i);
if (dev_info.isCompatible()) {
enable_device_id = i;
}
}
if (enable_device_id < 0) {
std::cerr << "GPU module isn't built for GPU" << std::endl;
return -1;
}
cv::cuda::setDevice(enable_device_id);
std::cout << "GPU is ready, device ID is " << num_devices << "\n";
// read one image
cv::Mat src_image = cv::imread("../111.jpg");
cv::Mat dst_image;
cv::cuda::GpuMat g_src_img(src_image);
cv::cuda::GpuMat g_dst_image;
cv::cuda::resize(g_src_img, g_dst_image, cv::Size(WIDTH, HEIGHT));
g_dst_image.download(dst_image);
cv::imshow("test", dst_image);
cv::waitKey(0);
cv::imwrite("test.jpg", dst_image);
return 0;
}
后来编译就出问题了,主要还是Eigen的问题
tensorflow-r1.14/tensorflow/contrib/makefile/downloads/eigen/unsupported/Eigen/CXX11/src/Tensor/TensorBlock.h: In static member function ‘static void Eigen::internal::TensorBlockIO<Scalar, StorageIndex, NumDims, Layout, BlockRead>::Copy(const Block&, StorageIndex, const Dimensions&, const Dimensions&, const Scalar*, Scalar*)’:
tensorflow-r1.14/tensorflow/contrib/makefile/downloads/eigen/unsupported/Eigen/CXX11/src/Tensor/TensorBlock.h:393:63: error: the value of ‘j’ is not usable in a constant expression
if (++block_iter_state[j].count < block_iter_state[j].size) {
^
tensorflow-r1.14/tensorflow/contrib/makefile/downloads/eigen/unsupported/Eigen/CXX11/src/Tensor/TensorBlock.h:392:16: note: ‘int j’ is not const
for (int j = 0; j < num_squeezed_dims; ++j) {
^
tensorflow-r1.14/tensorflow/contrib/makefile/downloads/eigen/unsupported/Eigen/CXX11/src/Tensor/TensorBlock.h:393:35: error: parse error in template argument list
if (++block_iter_state[j].count < block_iter_state[j].size) {
...
提取出主要问题
error: the value of ‘j’ is not usable in a constant expression
发现在github上也有类似问题https://github.com/tensorflow/tensorflow/issues/33280
但是并没有解决方案啊。。。
最后在LZ的仔细排查下最后发现了自己犯了一个很愚蠢的错误,在LZ的笔记本上有两个版本的OpenCV,
//这三行使用的是OpenCV4.2.0版本的
#include "opencv2/opencv.hpp"
#include "opencv2/core/cuda.hpp"
#include "opencv2/cudawarping.hpp"
//这个头文件使用的是OpenCV2.4.9版本的,这种肯定会出问题的,
#include "opencv2/gpu/gpu.hpp"
但是报的居然是tf中Eigen的问题,排查好久,差点以为要重新编译库了/(ㄒoㄒ)/~~
不知道github上那小哥解决问题了没有。。。