search

AlexeyAB / darknet

YOLOv4 / Scaled-YOLOv4 / YOLO - Neural Networks for Object Detection (Windows and Linux version of Darknet )
http://pjreddie.com/darknet/
Other
21.66k stars 7.96k forks source link

Segmentation fault with yolov3 continuous detection in python #3093

Open zrion opened 5 years ago

zrion commented 5 years ago

Hey there. I use original pythonv2 script along with the libso file to detect objects continuously in a loop, one image for each iteration. Loading net and meta seems fine, but the detection only runs for several iterations. In particular, I run Yolov3 tiny net on a GEFORCE GTX 1050Ti 4GB, and segmentation fault appears after ~168 iterations. At first I thought it may be memory leak issue, but since the detection function in the script already frees the image and detections at the end, it should not be the case. Single detection works fine. Any idea is appreciated!

LukeAI commented 5 years ago

I don't know but I reckon you are more likely to get help if you include the actual error message - does the same thing happen with other nets or just tiny?

primepake commented 5 years ago

I don't know but I reckon you are more likely to get help if you include the actual error message - does the same thing happen with other nets or just tiny?

I have same problem, some image I run segmentation fault (core dumped) image

some images are normal, I still don't understand it

primepake commented 5 years ago

sorry I wrong, in my case : name label is wrong tend to wrong in char on draw_detections_v3 fixed