Simplest QML algorithm to run on IBM QC

Hello, I am experimenting with pennylane’s QML algorithms such as variational classifiers (w keras).
All working correctly on pennylane simulator and also on ibmq_qasm_simulator.

However, when running on ibm’s quantum computers such as imbq_quito, due to the iterative nature of the variational classifier: it is launching too many jobs and I get long queues on ibmq. To the point it will take days before even getting metric results for the first epoch, even if I already run like 10/20 jobs.

I desperately need to get some full results (this week) of a single, simple ML classification algorithm using an actual IBM quantum computer.

What pennylane Demo algorithm and small, simple dataset do you suggest I use?
I need to minimize the jobs launched to ibmq and get some full results such as accuracy, even if not good.

Thank you very much in advance for any suggestions and help!

Hey @fersantanag!

This is the ideal case for using a runtime (see here for more details). I can’t guarantee that any of our demos will run for you in your timeframe, but our variational classifier demo might be a good one!

Hi @fersantanag, does this solution works for you? Please let us know, I have exactly same problem as yours, but using qiskit.ibmq.circuit_runner did not hep me.

@fersantanag I think you and @sdas may indeed have similar issues. You’ve commented on another thread that @sdas started (Running Pennylane circuit in IBMQ backend with Qiskit Runtime Session), so you’ll get notifications when new replies are made there as well.

Hopefully we can figure out what’s going on :slight_smile: