Description
We should allow seed as a tensor-valued argument to sample with intrinsic shape (), that is, a scalar.
The reason for doing this is to allow downstream components (executor cough cough) to allow broadcasting semantics over the seed.
This will allow users to do things like this
quantum_program.append_samplex_item(
template,
samplex,
samplex_arguments={seed: <1x20>, basis_changes.x: <9, 1>:})
in order to end up with 9 basis changes, but where the pauli randomizations are the same for each basis change.
Acceptance criteria
The seed is settable as a tensor-valued argument when invoking Samplex.sample().
Description
We should allow seed as a tensor-valued argument to sample with intrinsic shape (), that is, a scalar.
The reason for doing this is to allow downstream components (executor cough cough) to allow broadcasting semantics over the seed.
This will allow users to do things like this
in order to end up with 9 basis changes, but where the pauli randomizations are the same for each basis change.
Acceptance criteria
The seed is settable as a tensor-valued argument when invoking
Samplex.sample().