Hello, I have a few more questions about the runtime requirement.
Before testing phases, will there be a remote platform that can evaluate our model runtime on the target device?
Like it said in the organizer email, "on a recent smartphone CPU such model would be able to run in less than 0.2s...