AlexeyAB / darknet

YOLOv4 / Scaled-YOLOv4 / YOLO - Neural Networks for Object Detection (Windows and Linux version of Darknet )
http://pjreddie.com/darknet/
Other
21.73k stars 7.96k forks source link

Does Yolo works with pthreads instead of openmp #7213

Open ani16 opened 3 years ago

ani16 commented 3 years ago

Hello

I want to run yolo on hardware and it support only pthreads Library of C, not open mp..is it possible in yolo to run multi threads using pthreads not open mp library.

Thanks in advance

stephanecharette commented 3 years ago

Wouldn't you simply leave OPENMP=0 on line 6 of the makefile?

AlexeyAB commented 3 years ago

Run detection by using OpenCV: https://docs.opencv.org/master/da/d9d/tutorial_dnn_yolo.html the speed of detection will be very fast. or other frameworks: https://github.com/AlexeyAB/darknet#yolo-v4-in-other-frameworks

ani16 commented 3 years ago

Hi

Thanks for replying.

@stephanecharette: It will run the code with one core if I do openmp=0. I want to utilize all cores of a system.

@AlexeyAB : My hardware is very tighty coupled "free scale linux , 32-bit , 4 cores and 4 GB RAM", allowing only limited calls, not supporting OPENMP call. How can I accelerate detection on hardware without openmp? Does I really need pthread library to get all cores utilized or other method is also there?

"https://github.com/AlexeyAB/darknet#yolo-v4-in-other-frameworks"

I am implementing yolov3. Is there any method to accelerate detection on hardware for YOLOv3.

ani16 commented 3 years ago

Hi

I tried executing with one core on hardware..It throws an linking error.. actually Hardware is tightly coupled and don't support function __sqrt_finite.

Attached screenshot for reference. ![Uploading PXL_20210106_103840449.jpg…]()