Ask Your Question

OpenCV Yolo Tiny Slow

asked 2019-04-14 12:25:36 -0600

wpadka gravatar image

Hello Im trying to use my own trained model of yolov3-tiny in OpenCV. Im have code from this tutorial , but when im using videos or camera then this is going very slow(300ms) but when i`m using this same files cfg and weights in darknet then i have 40-50 fps. Is there any way to speed up this c++ code to get this same result, except lowering resolution in code, becouse it has normal speed only for 32x32 but then it detect nothing.

edit retag flag offensive close merge delete


Do you have the same results for original yolov3-tiny from ? I am experimenting with these models and object_detection.cpp example. Inference time is around 30ms for yolov3-tiny.

tobix10 gravatar imagetobix10 ( 2019-04-15 03:36:03 -0600 )edit

I guess you used GPU in darknet but CPU in openCV dnn

gino0717 gravatar imagegino0717 ( 2019-04-15 03:54:32 -0600 )edit

i needed change number of classes and filters and now this is working this same as mine

wpadka gravatar imagewpadka ( 2019-04-15 16:19:49 -0600 )edit

1 answer

Sort by ยป oldest newest most voted

answered 2019-04-15 04:26:48 -0600

Like stated by @gino0717, OpenCV DNN backend is not yet optimized with CUDA, which means it runs on CPU if you did not enable the OpenCL backend. However, there are optimized CPU models using the OpenVINO backend, have a look here:

edit flag offensive delete link more


this is only for Intel GPU what i see

wpadka gravatar imagewpadka ( 2019-04-15 16:18:58 -0600 )edit

yep that is possible, since they are acquired by Intel and probably Intel is pushing on that side. For the GSoC2019 program, there are students applying for integrating CUDA backends, so it might be slowly arriving ;)

StevenPuttemans gravatar imageStevenPuttemans ( 2019-04-16 02:40:44 -0600 )edit

Question Tools

1 follower


Asked: 2019-04-14 12:25:36 -0600

Seen: 2,181 times

Last updated: Apr 15 '19