How many video streams can single GPU handle for object detection

Question

I need to detect objects from multiple video streams at realtime (or close to it, like 10 FPS).

How many GPUs do I need to detect objects using YOLOv3 or MobileNet for, say, 10 video streams? Is it possible to use CPU or something else?

I don't need an exact number. I just need to understand scalability perspective and costs per single stream.

Answer 1

Since batch training is nothing else than computing multiple samples at once, you could test what the maximum batch size your GPU can handle for the model and that is how much you could calculate at once.

With regards to FPS it really depends on what GPU you have and how big the model is. Just try it out.