Fix inference_runner dynamic loading
- Update the inference_runner use case to work with the updated ML framework API
- Small change to setup script to only install dependencies for ExecuTorch-based use cases when we are setting up ExecuTorch resources
Change-Id: I517b39dccac1230ee86830c5dbe9274d883fa79c Signed-off-by: Alex Tawse alex.tawse@arm.com