I'm trying to run a PegasosQSVC on one of the IBMQ Quantum computers though the code just throws errors once we reach the training phase
ie: PegasosQSVC().fit
IBMQ.enable_account(token = 'MY TOKEN')
IBMQ.load_account()
IBMQ.providers()
provider = IBMQ.get_provider('ibm-q')
num_qubits = 2
Using these lines we get the least busy possible device as a backend
possible_devices = provider.backends(filters=lambda x: x.configuration().n_qubits >= num_qubits and x.configuration().simulator == False)
deviceBackend = least_busy(possible_devices)
Then we set up the QSVC
pegasos_backend = QuantumInstance(backend = deviceBackend)
feature_map = ZZFeatureMap(feature_dimension=2, reps=int(reps), entanglement=ent)
qkernel = QuantumKernel(feature_map=feature_map, quantum_instance=pegasos_backend)
pegasos_qsvc = PegasosQSVC(quantum_kernel=qkernel, C=int(C), num_steps=int(tau))
After
pegasos_qsvc.fit(train_features, train_labels)
It hangs then throws this error
FAILURE: job id:XXXXXXXXXXXX, status: 'FAIL_TO_GET_STATUS' Terra job error: '\' 530 Server Error....
The QSVC runs fine locally using
deviceBackend = BasicAer.get_backend('qasm_simulator')
I have also tried directly using the backend as a quantum instance though this makes the process hang
pegasos_backend = deviceBackend
Any help getting the QSVC to run on an IBM system would be greatly appreciated
