Salvatore Mandrà
Salvatore Mandrà
Python API is not working properly for complex arrays (see issue #31). There is also an issue in `benchmark/reference.cpp` when floating point are passed to `std::conj`. `std::conj` tries to convert...
Test fails for the python API when complex numbers are used. The culprit seems to be a missing parameter in `pythonAPI/hptt/hptt.py` at line 119 (`setConjA` is missing, which is instead...
For performance reasons, `simulate` may use either a single array of complex numbers (when using `optimize=evolution-einsum` or `backend='jax')` or a pair or real valued arrays (when using `optimize=evolution-hybridq`). For this...
At the moment, parameters for `cotengra` are not optimized for the input circuit. An autotuner would allow to select the right parameters to get the best contraction without the user's...
At the moment, QPU layouts are hard embedded in `hybridq.architecture`. To improve maintainability, use corresponding `API`s to retrieve the correct layout.
At the moment, most of the objects in **HybridQ** only implement `__str__`, while `__repr__` simply calls `__str__`. However, `__repr__` should return an unambiguous representation of the object while `__str__` should...