Merge branch 'master' into registers_feature

doc/code/qml.rst

@@ -6,4 +6,4 @@ qml
 .. automodapi:: pennylane
  :no-heading:
  :include-all-objects:
- :skip: Version, SimpleSpec, plugin_devices, plugin_converters, default_config, reload, version_info, defaultdict
+ :skip: plugin_converters, default_config, version_info, defaultdict

doc/development/deprecations.rst

@@ -9,6 +9,12 @@ deprecations are listed below.
 Pending deprecations
 --------------------
 
+* All of the legacy devices (any with the name ``default.qubit.{autograd,torch,tf,jax,legacy}``) are deprecated. Use ``default.qubit`` instead,
+ as it supports backpropagation for the many backends the legacy devices support.
+
+ - Deprecated in v0.38
+ - Will be removed in v0.39
+
 * The ``decomp_depth`` argument in ``qml.device`` is deprecated. 
 
  - Deprecated in v0.38

doc/releases/changelog-dev.md

@@ -21,9 +21,6 @@
  into multiple single-term measurements on a single tape, providing better support for simulators 
  [(#5987)](https://github.com/PennyLaneAI/pennylane/pull/5987)
 
- that can handle non-commuting observables but don't natively support multi-term observables.
- [(#5884)](https://github.com/PennyLaneAI/pennylane/pull/5884)
-
 * `SProd.terms` now flattens out the terms if the base is a multi-term observable.
  [(#5885)](https://github.com/PennyLaneAI/pennylane/pull/5885)
 
@@ -37,6 +34,9 @@
 
 <h3>Improvements 🛠</h3>
 
+* Added the `compute_sparse_matrix` method for `qml.ops.qubit.BasisStateProjector`.
+ [(#5790)](https://github.com/PennyLaneAI/pennylane/pull/5790)
+
 * `StateMP.process_state` defines rules in `cast_to_complex` for complex casting, avoiding a superfluous state vector copy in Lightning simulations
  [(#5995)](https://github.com/PennyLaneAI/pennylane/pull/5995)
 
@@ -60,6 +60,51 @@
 * Molecules and Hamiltonians can now be constructed for all the elements present in the periodic table.
  [(#5821)](https://github.com/PennyLaneAI/pennylane/pull/5821)
 
+* `qml.for_loop` and `qml.while_loop` now fallback to standard Python control
+ flow if `@qjit` is not present, allowing the same code to work with and without
+ `@qjit` without any rewrites.
+ [(#6014)](https://github.com/PennyLaneAI/pennylane/pull/6014)
+
+ ```python
+ dev = qml.device("lightning.qubit", wires=3)
+
+ @qml.qnode(dev)
+ def circuit(x, n):
+
+ @qml.for_loop(0, n, 1)
+ def init_state(i):
+ qml.Hadamard(wires=i)
+
+ init_state()
+
+ @qml.for_loop(0, n, 1)
+ def apply_operations(i, x):
+ qml.RX(x, wires=i)
+
+ @qml.for_loop(i + 1, n, 1)
+ def inner(j):
+ qml.CRY(x**2, [i, j])
+
+ inner()
+ return jnp.sin(x)
+
+ apply_operations(x)
+ return qml.probs()
+ ```
+
+ ```pycon
+ >>> print(qml.draw(circuit)(0.5, 3))
+ 0: ──H──RX(0.50)─╭●────────╭●──────────────────────────────────────┤ Probs
+ 1: ──H───────────╰RY(0.25)─│──────────RX(0.48)─╭●──────────────────┤ Probs
+ 2: ──H─────────────────────╰RY(0.25)───────────╰RY(0.23)──RX(0.46)─┤ Probs
+ >>> circuit(0.5, 3)
+ array([0.125 , 0.125 , 0.09949758, 0.15050242, 0.07594666,
+ 0.11917543, 0.08942104, 0.21545687])
+ >>> qml.qjit(circuit)(0.5, 3)
+ Array([0.125 , 0.125 , 0.09949758, 0.15050242, 0.07594666,
+ 0.11917543, 0.08942104, 0.21545687], dtype=float64)
+ ```
+
 * The `qubit_observable` function is modified to return an ascending wire order for molecular 
  Hamiltonians.
  [(#5950)](https://github.com/PennyLaneAI/pennylane/pull/5950)
@@ -136,6 +181,10 @@
  Instead, use `pennylane.gradients.classical_fisher` and `pennylane.gradients.quantum_fisher`.
  [(#5985)](https://github.com/PennyLaneAI/pennylane/pull/5985)
 
+* The legacy devices `default.qubit.{autograd,torch,tf,jax,legacy}` are deprecated.
+ Instead, use `default.qubit` as it now supports backpropagation through the several backends.
+ [(#5997)](https://github.com/PennyLaneAI/pennylane/pull/5997)
+
 <h3>Documentation 📝</h3>
 
 * Improves the docstring for `QuantumScript.expand` and `qml.tape.tape.expand_tape`.
@@ -177,10 +226,12 @@ Pietropaolo Frisoni,
 Emiliano Godinez,
 Austin Huang,
 Renke Huang,
+Josh Izaac,
 Soran Jahangiri,
 Christina Lee,
 William Maxwell,
 Vincent Michaud-Rioux,
+Anurav Modak,
 Mudit Pandey,
 Erik Schultheis,
 nate stemen.