Can I do a trade-off or benchmarking analysis between performance vs footprint for various AI hardware models?

This functionality is not supported at the moment but will be available soon.