yolo-tensorrt 运行 make报错 error: ‘int nvinfer1::MishPlugin marked ‘override’, but does not override

在机器上下载yolo-tensorrt项目，按照其测试步骤来操作，如下

在make这一步报错。

1. 完整报错信息

In file included from /home/aaeon/Desktop/yolov5-tensorrt/yolo-tensorrt/modules/trt_utils.h:38:0,
                 from /home/aaeon/Desktop/yolov5-tensorrt/yolo-tensorrt/modules/ds_image.h:28,
                 from /home/aaeon/Desktop/yolov5-tensorrt/yolo-tensorrt/modules/calibrator.h:29,
                 from /home/aaeon/Desktop/yolov5-tensorrt/yolo-tensorrt/modules/calibrator.cpp:26:
/home/aaeon/Desktop/yolov5-tensorrt/yolo-tensorrt/modules/mish.h:34:8: error: ‘int nvinfer1::MishPlugin::enq   ueue(int, const void* const*, void* const*, void*, cudaStream_t)’ marked ‘override’, but does not override
    int enqueue(int batchSize, const void* const* inputs, void* const* outputs, void* workspace,
        ^~~~~~~
In file included from /home/aaeon/Desktop/yolov5-tensorrt/yolo-tensorrt/modules/trt_utils.h:39:0,
                 from /home/aaeon/Desktop/yolov5-tensorrt/yolo-tensorrt/modules/ds_image.h:28,
                 from /home/aaeon/Desktop/yolov5-tensorrt/yolo-tensorrt/modules/calibrator.h:29,
                 from /home/aaeon/Desktop/yolov5-tensorrt/yolo-tensorrt/modules/calibrator.cpp:26:
/home/aaeon/Desktop/yolov5-tensorrt/yolo-tensorrt/modules/chunk.h:34:7: error: ‘int nvinfer1::Chunk::enqueue   (int, const void* const*, void* const*, void*, cudaStream_t)’ marked ‘override’, but does not override
   int enqueue(int batchSize, const void* const* inputs, void* const* outputs, void* workspace,
       ^~~~~~~
In file included from /home/aaeon/Desktop/yolov5-tensorrt/yolo-tensorrt/modules/trt_utils.h:40:0,
                 from /home/aaeon/Desktop/yolov5-tensorrt/yolo-tensorrt/modules/ds_image.h:28,
                 from /home/aaeon/Desktop/yolov5-tensorrt/yolo-tensorrt/modules/calibrator.h:29,
                 from /home/aaeon/Desktop/yolov5-tensorrt/yolo-tensorrt/modules/calibrator.cpp:26:
/home/aaeon/Desktop/yolov5-tensorrt/yolo-tensorrt/modules/hardswish.h:54:7: error: ‘int nvinfer1::Hardswish:   :enqueue(int, const void* const*, void* const*, void*, cudaStream_t)’ marked ‘override’, but does not overri   de
   int enqueue(int batchSize, const void* const* inputs, void* const* outputs, void* workspace,
       ^~~~~~~
In file included from /home/aaeon/Desktop/yolov5-tensorrt/yolo-tensorrt/modules/trt_utils.h:47:0,
                 from /home/aaeon/Desktop/yolov5-tensorrt/yolo-tensorrt/modules/ds_image.h:28,
                 from /home/aaeon/Desktop/yolov5-tensorrt/yolo-tensorrt/modules/calibrator.h:29,
                 from /home/aaeon/Desktop/yolov5-tensorrt/yolo-tensorrt/modules/calibrator.cpp:26:
/home/aaeon/Desktop/yolov5-tensorrt/yolo-tensorrt/modules/plugin_factory.h:139:7: error: ‘int nvinfer1::Yolo   Layer::enqueue(int, const void* const*, void* const*, void*, cudaStream_t)’ marked ‘override’, but does not    override
   int enqueue(int batchSize, const void* const* inputs, void* const* outputs, void* workspace,
       ^~~~~~~
CMakeFiles/detector.dir/build.make:62: recipe for target 'CMakeFiles/detector.dir/modules/calibrator.cpp.o'    failed
make[2]: *** [CMakeFiles/detector.dir/modules/calibrator.cpp.o] Error 1
CMakeFiles/Makefile2:77: recipe for target 'CMakeFiles/detector.dir/all' failed
make[1]: *** [CMakeFiles/detector.dir/all] Error 2
Makefile:83: recipe for target 'all' failed
make: *** [all] Error 2

根据报错信息，是tensorrt的插件出现问题。

2. 原因

具体来说，就是git clone得到的yolo-tensorrt版本为最新的，目前最新版yolo-tensorrt项目是基于tensorrt 8做的，tensorrt8 和 tensorrt7的部分插件不一致。

考虑我们的tensorrt版本是7。机器上tensorrt版本查询命令为

dpkg -l | grep TensorRT

我的查询结果如图

可以看到，安装的tensorrt版本为7.1.3。

3. 解决办法

或者安装tensorrt 8，或者使用tensorrt 7的yolo-tensortt，我选择后者。

tensorrt8和tensorrt7版本的yolo-tensorrt都在下述链接中，自取 。

链接：https://pan.baidu.com/s/1gF7AqCMkIxSrOfDFj0CLRQ
提取码：kdr7