tensorrtx icon indicating copy to clipboard operation
tensorrtx copied to clipboard

efficientnet make failed

Open stealth0414 opened this issue 2 years ago • 3 comments

Env

  • GPU, TX2
  • OS Ubuntu18.04
  • Cuda version 10.2
  • TensorRT version 8.2.1.8

About this repo

  • efficientnet
  • make

Your problem

When i make in build, it occurs: CMakeFiles/efficientnet.dir/build.make:62: recipe for target 'CMakeFiles/efficientnet.dir/efficientnet.cpp.o' failed make[2]: *** [CMakeFiles/efficientnet.dir/efficientnet.cpp.o] Error 1 CMakeFiles/Makefile2:67: recipe for target 'CMakeFiles/efficientnet.dir/all' failed make[1]: *** [CMakeFiles/efficientnet.dir/all] Error 2 Makefile:83: recipe for target 'all' failed make: *** [all] Error 2

stealth0414 avatar Jul 14 '22 07:07 stealth0414

Can you try tensorrt7.x? I don't think it support 8.x API.

wang-xinyu avatar Jul 18 '22 04:07 wang-xinyu

Thank you, tensorrt's version that comes with jetson nano, but I finally solved this problem. However I met a new problem when I convert my own trained .wts file to .engine sudo ./efficientnet -s ../model-421.wts model-421.engine b0 Loading weights: ../model-421.wts efficientnet: /home/huzhenyang/tensorrtx/efficientnet/utils.hpp:30: std::map<std::__cxx11::basic_string, nvinfer1::Weights> loadWeights(std::__cxx11::string): Assertion `input.is_open() && "Unable to load weight file."' failed. 已放弃 How can I fix this problem? :)

stealth0414 avatar Jul 18 '22 04:07 stealth0414

i met same question, could you tell me how to make efficientnet in tensorrt8.x successfully, thanks a lot.

DSPresearchlrj avatar Sep 15 '22 03:09 DSPresearchlrj

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

stale[bot] avatar Nov 14 '22 10:11 stale[bot]