Can opencv and Tensorflow use GPU simultaneously?
When I use Yolov3 in opencv with GPU for detection and use Deepsort in tensorflow with GPU for tracking simultaneously, I got the message of GPU out of memory. (I can use either of these two functions)
Have any idea to solve this problem, if I want to use these two function simultaneously? I have only one GPU.
I used to use KCF in opencv for tracking but it is slow and can only track same object, so I try to use Deepsort in tensorflow with GPU.
If there is no sollution for the problem, whether there is recommendation of another fast tracking algorithem for multiple object like DeepSort?
What is that you don't understand?
Out of memory means that your GPU doesn't have enough memory to load both networks.
@Sshwan what GPU you have? You can reduce memory consumption by using the
DNN_TARGET_CUDA_FP16target. You will reduce the memory footprint by approximately 2x.