When connecting to the internet, and using the OPENVINO model, is this technically cloud computing?
No service use
michaelsalerno
No, cloud computing means you are running inference on some other machine in the cloud. In the case of depthai, the model is loaded on the OAK device and is running on RVC VPU. This is called edge computing.
Thanks,
Jaka
Thank you, I really appreciate it!
If it is still only working through wifi, does that mean its edge computing still? Or does this mean that it is cloud computing?
michaelsalerno
Just because the OAK device is connected via WiFi does not mean it’s cloud computing. WiFi is simply the medium for network connectivity. If the inference is still happening locally on the OAK device and not on a cloud server, it’s still edge computing.
Cloud computing only applies if the actual computation (inference) is happening on a remote cloud server.
Thanks,
Jaka
Thank you for all your help!
One last question,
What NN is it using? Luxonis published information on latency that I would like to use, and I am curious to know which one it is so I can accurately talk about latency.
luxonis/depthai-experimentstree/master/gen2-people-tracker
The model that is using this.
From Michael
michaelsalerno
As readme suggests, it's mobilenetV2 backbone. We don't have latency tests for it. DIY
Thanks,
Jaka
Is there a simple way to test that?
michaelsalerno
Run the script with DEPTHAI_LEVEL=TRACE then check the latencies for NN.
Thanks,
Jaka
Should I put that in the main.py code for it to pop up while i am recording?
I am just a little confused on the where to apply the DEPTHAI_LEVEL=TRACE
michaelsalerno
Explained here
Thanks,
Jaka