Closed lusi1990 closed 3 years ago
Hi, 再仔细看下readme中int8的说明,需要准备calibration data
Hi, 再仔细看下readme中int8的说明,需要准备calibration data
感谢, 已经解决了
你好 是哪里有问题呢?
make [ 20%] Building NVCC (Device) object CMakeFiles/myplugins.dir/myplugins_generated_yololayer.cu.o tensorrtx/yolov4/yololayer.h(105): warning: function "nvinfer1::IPluginV2Ext::configurePlugin(const nvinfer1::Dims , int32_t, const nvinfer1::Dims , int32_t, const nvinfer1::DataType , const nvinfer1::DataType , const __nv_bool , const __nv_bool , nvinfer1::PluginFormat, int32_t)" is hidden by "nvinfer1::YoloLayerPlugin::configurePlugin" -- virtual function override intended?
/tensorrtx/yolov4/yololayer.h(105): warning: function "nvinfer1::IPluginV2Ext::configurePlugin(const nvinfer1::Dims , int32_t, const nvinfer1::Dims , int32_t, const nvinfer1::DataType , const nvinfer1::DataType , const bool , const bool , nvinfer1::PluginFormat, int32_t)" is hidden by "nvinfer1::YoloLayerPlugin::configurePlugin" -- virtual function override intended?
[ 40%] Building NVCC (Device) object CMakeFiles/myplugins.dir/myplugins_generated_mish.cu.o tensorrtx/yolov4/mish.h(62): warning: function "nvinfer1::IPluginV2Ext::configurePlugin(const nvinfer1::Dims , int32_t, const nvinfer1::Dims ,/home/keyvalue int32_t, const nvinfer1::DataType , const nvinfer1::DataType , const __nv_bool , const __nv_bool , nvinfer1::PluginFormat, int32_t)" is hidden by "nvinfer1::MishPlugin::configurePlugin" -- virtual function override intended?
tensorrtx/yolov4/mish.h(62): warning: function "nvinfer1::IPluginV2Ext::configurePlugin(const nvinfer1::Dims , int32_t, const nvinfer1::Dims , int32_t, const nvinfer1::DataType , const nvinfer1::DataType , const bool , const bool , nvinfer1::PluginFormat, int32_t)" is hidden by "nvinfer1::MishPlugin::configurePlugin" -- virtual function override intended?
Scanning dependencies of target myplugins [ 60%] Linking CXX shared library libmyplugins.so [ 60%] Built target myplugins Scanning dependencies of target yolov4 [ 80%] Building CXX object CMakeFiles/yolov4.dir/yolov4.cpp.o [100%] Linking CXX executable yolov4 [100%] Built target yolov4 Then /tensorrtx/yolov4/build# ./yolov4 -s Loading weights: ../yolov4.wts CUDA error 222 at /tensorrtx/yolov4/yololayer.cu:19yolov4: /tensorrtx/yolov4/yololayer.cu:19: nvinfer1::YoloLayerPlugin::YoloLayerPlugin(): Assertion `0' failed. Aborted (core dumped)
Hi, 再仔细看下readme中int8的说明,需要准备calibration data
感谢, 已经解决了
感谢,感谢
CUDA: 10.2 CUDNN: 8.0.5 tensorRT: 7.2.2.3 OS: Ubuntu 18.04
yolov5 的版本是 4.0
cmake 输出
make 输出
./yolov5 -s
FP32 和 FP16 都能成功执行并且推断. opencv 是自己编译安装的. 版本是最新的