YOLOv3) compiles and run. But no predictions

Question

I'm using macOS Catalina.

I'm just following the basic tutorial which are written in the Darknet page.

when I run the command

$make
$./darknet detect cfg/yolov3.cfg yolov3.weights data/dog.jpg

then the Yolo is not predict after just showing the running time

Loading weights from yolov3.weights...Done!
data/dog.jpg: Predicted in 19.215062 seconds.

Also, prediction.jpg is saved without any boundary box

I tried also

$./darknet detect cfg/yolov3.cfg yolov3.weights data/dog.jpg -thresh 0

But it doesn't work. Can anyone help me please?

https://datascience.stackexchange.com/ seems to be more appropriate for this topic. — Kate Melnykova, May 14 '20 at 07:03

score 0 · Accepted Answer · answered May 14 '20 at 06:58

0

When you run detect command, the prediction is saved as an image instead of showing any log. Look into the current data folder, there should be an image named predictions.jpg. This image contains all the bounding boxes with classes.

If you want to print your predictions, you can use darknet.py file inside darknet folder.

answered May 14 '20 at 06:58

Zabir Al Nazi

10,298
4
33
60

Thank you for replying! However, the predictions.jpg is not containg bounding boxes and classes. – TaeYoon Kim May 14 '20 at 07:05
That means your model didn't predict any objects in the image, try with some other image. – Zabir Al Nazi May 14 '20 at 07:11
I've tried all the sample images provided from them and decrease thresh to 0, but it doesn't work. – TaeYoon Kim May 14 '20 at 07:27
You are right! I reinstalled the weight. and it works! I never thought about the weight file because it was loaded successfully. Thank you very much:) – TaeYoon Kim May 14 '20 at 08:14

YOLOv3) compiles and run. But no predictions

1 Answers1