Ask Your Question

Revision history [back]

click to hide/show revision 1
initial version

Difference in time for YOLOv3

Hi, I run dnn/object_detection.cpp with params: -c=yolov3.cfg -m=yolov3.weights -classes=coco.names --scale=0.00392 --width=416 --height=416 and it works, but time it takes for one image is about 4 seconds. Although in the image on top is written about 745 ms. I run in release and use opencv4. C:\fakepath\Screenshot1.png

Also in this link (https://github.com/opencv/opencv/pull/11322) developers writes that median time per 416x416 image is 216.11ms.

What may be the reason of this time difference?

I have Ubuntu OS: C:\fakepath\Screenshot2.png

click to hide/show revision 2
None

updated 2018-09-28 06:13:37 -0600

berak gravatar image

Difference in time for YOLOv3

Hi, I run dnn/object_detection.cpp with params: -c=yolov3.cfg -m=yolov3.weights -classes=coco.names --scale=0.00392 --width=416 --height=416 and it works, but time it takes for one image is about 4 seconds. Although in the image on top is written about 745 ms. I run in release and use opencv4. C:\fakepath\Screenshot1.pngopencv4.

Also in this link (https://github.com/opencv/opencv/pull/11322) developers writes that median time per 416x416 image is 216.11ms.

What may be the reason of this time difference?

I have Ubuntu OS: C:\fakepath\Screenshot2.pngOS:

Difference in time for YOLOv3

Hi, I run dnn/object_detection.cpp with params: -c=yolov3.cfg -m=yolov3.weights -classes=coco.names --scale=0.00392 --width=416 --height=416 and it works, but time it takes for one image is about 4 seconds. Function: net.forward(outs, getOutputsNames(net)) takes the most time. Although in the image (C:\fakepath\Screenshot1.png) on top is written about 745 ms. I run in release and use opencv4.

Also in this link (https://github.com/opencv/opencv/pull/11322) developers writes that median time per 416x416 image is 216.11ms.

What may be the reason of this time difference?

I have Ubuntu OS:OS: C:\fakepath\Screenshot2.png

Difference in time for YOLOv3

Hi, I run dnn/object_detection.cpp with params: -c=yolov3.cfg -m=yolov3.weights -classes=coco.names --scale=0.00392 --width=416 --height=416 and it works, but time it takes for one image is about 4 seconds. Function: net.forward(outs, getOutputsNames(net)) takes the most time. Although in on the image (C:\fakepath\Screenshot1.png) (https://drive.google.com/open?id=1kZIUS5Dct3sPRZrMKsixFCWLgizbKnzd) on top is written about 745 ms. I run in release and use opencv4.

Also in this link (https://github.com/opencv/opencv/pull/11322) developers writes that median time per 416x416 image is 216.11ms.

What may be the reason of this time difference?

I have Ubuntu OS: C:\fakepath\Screenshot2.pngOS: https://drive.google.com/open?id=1MvbLah4gNFB-P2JvW5P5MZ7znhlX5u8Y