antalszava
antalszava
### Feature details When passing in the wires to `qml.SparseHamiltonian`, one could check that the shape of the sparse Hamiltonian corresponds to the number of wires passed in. ### Implementation...
**Context:** ```py pl_tensor = qml.numpy.array(0.3, requires_grad=True) torch_tensor = torch.tensor(0.3, requires_grad=True) qml.math.allclose(pl_tensor, torch_tensor) ``` Raises ```py RuntimeError: Can't call numpy() on Tensor that requires grad. Use tensor.detach().numpy() instead. ``` **Description of...
### Feature details The #2110 PR is adding a custom `InterfaceUnsupportedError` that is being used with features unsupported by the JAX interface. It would come as a next step to:...
### Feature details We have a leftover `supports_reversible_diff` key in device capabilities. This is a remnant of an earlier implementation and can be removed from the code base. The logic...
### Feature details Any keyword argument can be passed to a QNode or its qnode decorator without errors arising. This can result in unexpected behaviour if users make a typo...
### Expected behavior Using `default.qubit.torch` and backpropagation, `return qml.probs(0), qml.probs(1)` returns a `(2, 2)` shape tensor (just as with TensorFlow). ### Actual behavior A `(4,)` tensor is returned for `return...
### Feature details We'd like to have a single `default.qubit` device by merging each of the subclasses of `default.qubit`. The new `default.qubit` device should use `qml.math` such that it provides...
### Feature details Add two new transforms: * `qml.transforms.quantum_jacobian` and * `qml.transforms.natural_jacobian`. This would be a continuation of https://github.com/PennyLaneAI/pennylane/pull/1259. ### Implementation _No response_ ### How important would you say this...
### Feature details It would be good to add `doctest` support. See https://github.com/PennyLaneAI/pennylane/pull/705. ### Implementation _No response_ ### How important would you say this feature is? 1: Not important. Would...
### Feature details Interfaces in PennyLane allow using Machine Learning frameworks. Each interface is implemented as a single component in PennyLane that takes on two responsibilities: 1. Convert between NumPy...