TwirlingOptions

class TwirlingOptions(*args, **kwargs)

GitHub

Twirling options.

Attributes

enable_gates

Type: UnsetType | bool

Default value: Unset

Whether to apply 2-qubit Clifford gate twirling. Default: False.

enable_measure

Type: UnsetType | bool

Default value: Unset

Whether to enable twirling of measurements.

Twirling will only be applied to those measurement registers not involved within a conditional logic. Default: True for Estimator, false for Sampler.

num_randomizations

Type: UnsetType | int | Literal['auto']

Default value: Unset

The number of random samples to use when twirling or peforming sampled mitigation.

If num_randomizations is “auto”, for every pub executed shots times:

If shots_per_randomization is also “auto”, shots_per_randomization is set first as described below, then num_randomizations is set as ceil(shots/shots_per_randomization), where ceil is the ceiling function.

Otherwise, the value is set to ceil(shots/shots_per_randomization).

Default: “auto”.

Note

The shots value specified in a PUB or in the run() method is considered part of the primitive execution interface and therefore is always obeyed. default_shots, on the other hand, is considered a Qiskit Runtime specific option. Therefore, the product of num_randomizations and shots_per_randomization takes precedence over default_shots.

shots_per_randomization

Type: UnsetType | int | Literal['auto']

Default value: Unset

The number of shots to run for each random sample.

If “auto”, for every pub executed shots times:

If num_randomizations is also “auto”, the value is set to 64 for PEC mitigation or to max(64, ceil(shots / 32)) in all other cases, where ceil is the ceiling function.

Otherwise, the value is set to ceil(shots/num_randomizations).

Default: “auto”.

Note

strategy

Type: UnsetType | Literal['active', 'active-accum', 'active-circuit', 'all']

Default value: Unset

Specify the strategy of twirling qubits in identified layers of 2-qubit twirled gates.

Allowed values are:

If "active" only the instruction qubits in each individual twirled layer will be twirled.

If "active-circuit" the union of all instruction qubits in the circuit will be twirled in each twirled layer.

If "active-accum" the union of instructions qubits in the circuit up to the current twirled layer will be twirled in each individual twirled layer.

If "all" all qubits in the input circuit will be twirled in each twirled layer.

Default: “active-accum”.

Methods

Was this page helpful?

Report a bug or request content on GitHub.