Add EstimatorV2 #2088
Add EstimatorV2 #2088
Conversation
|
Could you summarize how it's determined whether to use true sampling, or to use a normal distribution approximation? What happens if the user doesn't specify anything? |
|
@kevinsung I simplified because it is confusing and true sampling was deprecated in V1. As documented, there is only a normal distribution approximation |
| ).result() | ||
|
|
||
| # calculate expectation values (evs) and standard errors (stds) | ||
| rng = np.random.default_rng(self.options.run_options.get("seed_simulator")) |
There was a problem hiding this comment.
I'm wondering that rng should be shared in a _run() call.
There was a problem hiding this comment.
I think implementation of rng is as same as that of qiskit.primitives https://github.com/Qiskit/qiskit/blob/090b2b193d6de88017e26d99c0ac379b85382395/qiskit/primitives/statevector_estimator.py#L142
There was a problem hiding this comment.
Yes. This is the point where we decided to make this specification for maintainability when we implemented SamplerV2.
| return PrimitiveResult([self._run_pub(pub) for pub in pubs]) | ||
|
|
||
| def _run_pub(self, pub: EstimatorPub) -> PubResult: | ||
| circuit = pub.circuit.copy() |
There was a problem hiding this comment.
I think pub.circuit should be compiled to AerCircuit here to avoid copy. This will be an optimization in next releases.
doichanj
added
enhancement
Summary
Details and comments