TensoredMeasFitter gives "TensoredMeasFitter doesn't support subset_fitter." error

[omarshehab]

Informations

• Qiskit Ignis version:0.6.0
• Python version:3.9.5 (default, May 18 2021, 12:31:01) [Clang 10.0.0 ]
• Operating system:Darwin

What is the current behavior?

When I use TensoredMeasFitter for a 15 qubit QSVM experiment, I get the following error.

QiskitError                               Traceback (most recent call last)
/var/folders/5p/_ydvdq5959v8p0kntv58swdc0000gn/T/ipykernel_2969/2322411278.py in <module>
----> 1 adhoc_matrix_train_q_em = adhoc_kernel.evaluate(x_vec=train_features)
      2 adhoc_matrix_test_q_em = adhoc_kernel.evaluate(x_vec=test_features,
      3                                           y_vec=train_features)

~/opt/anaconda3/envs/py39/lib/python3.9/site-packages/qiskit_machine_learning/kernels/quantum_kernel.py in evaluate(self, x_vec, y_vec)
    331                 ]
    332 
--> 333                 results = self._quantum_instance.execute(circuits)
    334 
    335                 matrix_elements = [

~/opt/anaconda3/envs/py39/lib/python3.9/site-packages/qiskit/utils/quantum_instance.py in execute(self, circuits, had_transpiled)
    654                         tmp_fitter = meas_error_mitigation_fitter.subset_fitter(curr_qubit_index)
    655                     else:
--> 656                         raise QiskitError(
    657                             "{} doesn't support subset_fitter.".format(
    658                                 meas_error_mitigation_fitter.__class__.__name__

QiskitError: "TensoredMeasFitter doesn't support subset_fitter."

Steps to reproduce the problem

Instantiate QuantumInstance in the following manner and use it in a QSVM project with 15 qubits.

adhoc_backend = QuantumInstance(backend, shots=1024,
                                seed_transpiler=seed, optimization_level = 3, measurement_error_mitigation_cls=
TensoredMeasFitter)

What is the expected behavior?

It should work without any error.

Suggested solutions