What if I need to use a custom face detection model such as those based on SSD or Yolo face. Could you make a blog post for yolov3?

This is a good implementation but won’t it take more computation burden if run on webcams?

Thanks for such a beautiful explanation.