fatal error: NvInfer.h: No such file or directory | TensorRT 报错处理

本文介绍了在使用TensorRT时遇到的两个常见编译错误：找不到NvInfer.h头文件和log方法重定义问题。解决方案包括在CMakeLists.txt中正确配置TensorRT路径，并在遇到log方法冲突时考虑回退到CUDA 10 + TensorRT 7的组合。同时，提到了其他可能遇到的库文件找不到的问题，如libmyelin.so.1和libcudnn.so.7。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

fatal error: NvInfer.h: No such file or directory | 报错处理

9-9
1-1

💜 遇到报错【 fatal error: NvInfer.h: No such file or directory】：

fatal error: NvInfer.h: No such file or directory

#include "NvInfer.h"

通过 find -name NvInfer.h 查找发现该头文件所在目录

可以发现，该头文件是属于 TensorRT 下的一个专有头文件，因此我们的C++ 代码在编译时需要能够找到它

find -name NvInfer.h

./project/project21/modelTrans/tensorRT/tensorRT7/TensorRT-7.0.0.11/include/NvInfer.h
./project/project21/modelTrans/tensorRT/tensorRT7/TensorRT-7.1.3.4/include/NvInfer.h
./project/project21/modelTrans/tensorRT/tensorRt8/TensorRT-8.0.0.3/include/NvInfer.h
./project/project21/modelTrans/transPro/yolo-tensorrt/modules/NvInfer.h

❤️ 解决方法

在 CMakeLists.txt 中配置 TensorRT-8


# tensorRT
include_directories(/home/墨理/project/project21/modelTrans/tensorRT/tensorRt8/TensorRT-8.0.0.3/include/)
link_directories(/home/墨理/project/project21/modelTrans/tensorRT/tensorRt8/TensorRT-8.0.0.3/lib/)

💜 遇到报错【NvInferRuntimeCommon.h:1203:18: error: overriding ‘virtual void nvinfer1::ILogger::log】：

crnn/logging.h 下的 log() 方法和 NvInferRuntimeCommon.h 中的 log（）方法重复定义

/home/墨理/project/project21Next/modelTrans/tvm/tensorrtx_Cuda11/crnn/logging.h:239:10: error: looser throw specifier for ‘virtual void Logger::log(nvinfer1::ILogger::Severity, const char*)’
     void log(Severity severity, const char* msg) override

...

TensorRT-8.0.0.3/include/NvInferRuntimeCommon.h:1203:18: error:   overriding ‘virtual void nvinfer1::ILogger::log(nvinfer1::ILogger::Severity, const AsciiChar*) noexcept’
     virtual void log(Severity severity, AsciiChar const* msg) noexcept = 0;

观察 crnn/logging.h 发现

    void log(Severity severity, const char* msg) override
    {
        LogStreamConsumer(mReportableSeverity, severity) << "[TRT] " << std::string(msg) << std::endl;
    }

❤️ 解决方法，使用 cuda-10 + tensorRT7 的组合，该冲突即消失


# cuda
include_directories(/usr/local/cuda-10.0/include)
link_directories(/usr/local/cuda-10.0/lib64)


# tensorRT
#include_directories(/home/墨理/project/project21/modelTrans/tensorRT/tensorRt8/TensorRT-8.0.0.3/include/)
#link_directories(/home/墨理/project/project21/modelTrans/tensorRT/tensorRt8/TensorRT-8.0.0.3/lib/)

# tensorRT 7
include_directories(/home/墨理/project/project21/modelTrans/tensorRT/tensorRT7/TensorRT-7.0.0.11/include/)
link_directories(/home/墨理/project/project21/modelTrans/tensorRT/tensorRT7/TensorRT-7.0.0.11/lib/)