search

AlexeyAB / darknet

YOLOv4 / Scaled-YOLOv4 / YOLO - Neural Networks for Object Detection (Windows and Linux version of Darknet )
http://pjreddie.com/darknet/
Other
21.75k stars 7.96k forks source link

Darknet stop immediatly without any error message #8369

Open kirin111697 opened 2 years ago

kirin111697 commented 2 years ago

The problem I encounter is that my darknet build seems like not doing anything. I am trying to train my own customize model using darknet. After I successfully build it, it doesn't seem to do anything on any command. image image Every command seemed to end immediately and didn't print any error message. I build the executable using cmake on windows10 with opencv4. I am totally at lost here about what the problem might be. Any help or advice is appreciated.

stephanecharette commented 2 years ago

1) -clear 1 is not a valid parameter. You probably meant to use -clear. But the way you are using it in comment above is also incorrect. 2) See: https://www.ccoderun.ca/programming/darknet_faq/#training_command 3) Are you certain that all your paths and everything is 100% correct in your obj.data file? 4) Are you certain you checked out the right repo and built it correctly? (e.g., https://www.ccoderun.ca/programming/darknet_faq/#how_to_build_on_windows) 5) Is there anything in the Windows system log or application log that might indicate why darknet.exe is exiting immediately? Maybe a bad or missing opencv or cuda DLL?

kirin111697 commented 2 years ago

Hi, thanks for the reply. I think my problem here is on my darknet since it doesn't even responding to the provided commands on the darknet github page. I am certain that my obj.data file is correct and I am using the right version. I checked the logs and there aren't anything alarming. image inside data folder image the darknet exe file is in the root folder