编译lite.ai.toolkit后运行yolo5代码出现异常 - Githubissues

DefTruth / lite.ai.toolkit

🛠 A lite C++ toolkit of awesome AI models, support ONNXRuntime, MNN, TNN, NCNN and TensorRT.

https://github.com/DefTruth/lite.ai.toolkit

GNU General Public License v3.0

3.59k stars 684 forks source link

编译lite.ai.toolkit后运行yolo5代码出现异常 #342

Closed l7518597 closed 2 years ago

l7518597 commented 2 years ago

iss

l7518597 commented 2 years ago

求助

DefTruth commented 2 years ago

可以参考一下这个issue:

340

l7518597 commented 2 years ago

好的谢谢

l7518597 commented 2 years ago

试过了换了opencv4.5.2 还是同样的情况

l7518597 commented 2 years ago

DefTruth commented 2 years ago

可以看下模型文件的路径等是否正常，我写的examples里是需要从指定路径读取模型的，你看下test_lite_yolov5_v6.0.cpp中指定的模型是否存在？不存在的话需要从README.md中提到的地址进行下载

l7518597 commented 2 years ago

我自己训练了模型，再准备的图片可以跑了。用自带的不行。感谢！！

l7518597 commented 2 years ago

另外资询个问题: 我编译0.1.1版本可以但是直接github克隆的编译不了是因为克隆版里cmake内的opencv.cmake和0.1.1里的setup_opencv_libs.cmake设置问题吗

l7518597 commented 2 years ago

imread的参数是图片路径,有没有办法可以是传图片的字节集和长度来加载图片