Added support for EstimatorV2 primitives #48
Conversation
OkuyanBoga
requested review from
smens,
edoaltamura and
oscar-wallis
as code owners
There was a problem hiding this comment.
@edoaltamura @oscar-wallis please do corrections according to my reviews.
qiskit_machine_learning/gradients/param_shift/param_shift_estimator_gradient.py
Show resolved
Hide resolved
| ) | ||
| results = job.result() | ||
| elif isinstance(self._estimator, BaseEstimatorV2): | ||
| job = self._estimator.run(PUBs, precision=0.001, **options) |
There was a problem hiding this comment.
precision parameter might introduce issues for tests, we need to think about how to make stable tests for V2.
There was a problem hiding this comment.
Qiskit uses a default value of 0.0: https://github.com/Qiskit/qiskit/blob/8ccbc8d28f34e3f2e1e0e8cbb66c4da95754b341/qiskit/primitives/statevector_estimator.py#L113
So to make tests reproducible we could fetch
precision=self._default_precisionSince we are already within the BaseEstimatorV2 case, we can assume that _default_precision is defined in the superclass.
qiskit_machine_learning/gradients/param_shift/param_shift_estimator_gradient.py
Show resolved
Hide resolved
qiskit_machine_learning/gradients/spsa/spsa_estimator_gradient.py
Outdated
Show resolved
Hide resolved
qiskit_machine_learning/gradients/spsa/spsa_estimator_gradient.py
Outdated
Show resolved
Hide resolved
| self.num_qubits = circuit.num_qubits | ||
| print('Warning: No number of qubits provided, using it from provided circuit, if the circuit is transpiled this can be wrong and cause issues.') |
There was a problem hiding this comment.
This is required since we might lose information about actual number of qubits of the circuit after the transpilation, not sure if it can simply be retrieved from transpiled_circuit.
qiskit_machine_learning/gradients/param_shift/param_shift_estimator_gradient.py
Outdated
Show resolved
Hide resolved
qiskit_machine_learning/gradients/param_shift/param_shift_estimator_gradient.py
Outdated
Show resolved
Hide resolved
qiskit_machine_learning/gradients/param_shift/param_shift_estimator_gradient.py
Outdated
Show resolved
Hide resolved
qiskit_machine_learning/gradients/spsa/spsa_estimator_gradient.py
Outdated
Show resolved
Hide resolved
Co-authored-by: Edoardo Altamura <38359901+edoaltamura@users.noreply.github.com>
Co-authored-by: Edoardo Altamura <38359901+edoaltamura@users.noreply.github.com>
Co-authored-by: Edoardo Altamura <38359901+edoaltamura@users.noreply.github.com>
Co-authored-by: Edoardo Altamura <38359901+edoaltamura@users.noreply.github.com>
…mator_gradient.py Co-authored-by: Edoardo Altamura <38359901+edoaltamura@users.noreply.github.com>
Co-authored-by: Edoardo Altamura <38359901+edoaltamura@users.noreply.github.com>
…mator_gradient.py Co-authored-by: Edoardo Altamura <38359901+edoaltamura@users.noreply.github.com>
Co-authored-by: Edoardo Altamura <38359901+edoaltamura@users.noreply.github.com>
Co-authored-by: Edoardo Altamura <38359901+edoaltamura@users.noreply.github.com>
…mator_gradient.py Co-authored-by: Edoardo Altamura <38359901+edoaltamura@users.noreply.github.com>
| @@ -23,6 +23,9 @@ | |||
| from qiskit_machine_learning.circuit.library import QNNCircuit | |||
|
|
|||
| from qiskit_machine_learning.neural_networks.estimator_qnn import EstimatorQNN | |||
| from qiskit_ibm_runtime import Session, EstimatorV2 | |||
| from qiskit.providers.fake_provider import GenericBackendV2 | |||
| from qiskit.transpiler.preset_passmanagers import generate_preset_pass_manager | |||
|
|
|||
| CASE_DATA = { | |||
There was a problem hiding this comment.
@oscar-wallis Can you please copy paste this as CASE_DATA_V2 and adapt for V2?
| self, | ||
| TestCase, | ||
| ): | ||
| self.backend = GenericBackendV2(num_qubits=8) |
There was a problem hiding this comment.
Change this to num_qubits=3 or 4
Added support for V2 primitives for estimator_qnn and gradients.
TO DO: