Skip to content

Commit

Permalink
some updates to adjoint Jacobian tests
Browse files Browse the repository at this point in the history
  • Loading branch information
@AmintorDusko
AmintorDusko committed Oct 2, 2023
1 parent 0997c9d commit 2ed8493
Showing 1 changed file with 110 additions and 112 deletions.
222 changes: 110 additions & 112 deletions tests/test_adjoint_jacobian.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -124,9 +124,7 @@ def test_not_expval(self, dev):
with pytest.raises(
qml.QuantumFunctionError, match="Adjoint differentiation method does not"
):
# self.process_and_compute_derivatives(dev, tape)
dev.supports_derivatives(AdjointConfig, circuit=tape)
# dev.adjoint_jacobian(tape)

with qml.tape.QuantumTape() as tape:
qml.RX(0.1, wires=0)
Expand All @@ -143,7 +141,6 @@ def test_not_expval(self, dev):
match=message,
):
dev.supports_derivatives(AdjointConfig, circuit=tape)
# dev.adjoint_jacobian(tape)

def test_finite_shots_warns(self):
"""Tests warning raised when finite shots specified"""
Expand All @@ -156,7 +153,6 @@ def test_finite_shots_warns(self):
with pytest.warns(
UserWarning, match="Requested adjoint differentiation to be computed with finite shots."
):
# dev.adjoint_jacobian(tape)
dev.supports_derivatives(AdjointConfig, circuit=tape)

def test_empty_measurements(self, dev):
Expand All @@ -180,7 +176,6 @@ def test_unsupported_op(self, dev):
with pytest.raises(
qml.QuantumFunctionError, match="The CRot operation is not supported using the"
):
# dev.adjoint_jacobian(tape)
dev.supports_derivatives(AdjointConfig, circuit=tape)

@pytest.mark.skipif(not ld._CPP_BINARY_AVAILABLE, reason="Lightning binary required")
Expand All @@ -193,7 +188,6 @@ def test_proj_unsupported(self, dev):
with pytest.raises(
qml.QuantumFunctionError, match="differentiation method does not support the Projector"
):
# dev.adjoint_jacobian(tape)
dev.supports_derivatives(AdjointConfig, circuit=tape)

with qml.tape.QuantumTape() as tape:
Expand All @@ -203,7 +197,6 @@ def test_proj_unsupported(self, dev):
with pytest.raises(
qml.QuantumFunctionError, match="differentiation method does not support the Projector"
):
# dev.adjoint_jacobian(tape)
dev.supports_derivatives(AdjointConfig, circuit=tape)

@pytest.mark.parametrize("theta", np.linspace(-2 * np.pi, 2 * np.pi, 7))
Expand Down Expand Up @@ -259,14 +252,14 @@ def test_ry_gradient(self, par, tol, dev):
qml.RY(par, wires=[0])
qml.expval(qml.PauliX(0))

tape.trainable_params = {0}
trainable_params = {0}

# gradients
exact = np.cos(par)
grad_A = dev.adjoint_jacobian(tape)
calculated_val = self.process_and_compute_derivatives(dev, tape, trainable_params)

# different methods must agree
assert np.allclose(grad_A, exact, atol=tol, rtol=0)
assert np.allclose(calculated_val, exact, atol=tol, rtol=0)

def test_rx_gradient(self, tol, dev):
"""Test that the gradient of the RX gate matches the known formula."""
Expand All @@ -276,10 +269,10 @@ def test_rx_gradient(self, tol, dev):
qml.RX(a, wires=0)
qml.expval(qml.PauliZ(0))

# circuit jacobians
dev_jacobian = dev.adjoint_jacobian(tape)
expected_jacobian = -np.sin(a)
assert np.allclose(dev_jacobian, expected_jacobian, atol=tol, rtol=0)
# circuit Jacobians
calculated_val = self.process_and_compute_derivatives(dev, tape)
expected_val = -np.sin(a)
assert np.allclose(calculated_val, expected_val, atol=tol, rtol=0)

def test_multiple_rx_gradient_pauliz(self, tol, dev):
"""Tests that the gradient of multiple RX gates in a circuit yields the correct result."""
Expand All @@ -293,10 +286,10 @@ def test_multiple_rx_gradient_pauliz(self, tol, dev):
for idx in range(3):
qml.expval(qml.PauliZ(idx))

# circuit jacobians
dev_jacobian = dev.adjoint_jacobian(tape)
expected_jacobian = -np.diag(np.sin(params))
assert np.allclose(dev_jacobian, expected_jacobian, atol=tol, rtol=0)
# circuit Jacobians
calculated_val = self.process_and_compute_derivatives(dev, tape)
expected_val = -np.diag(np.sin(params))
assert np.allclose(calculated_val, expected_val, atol=tol, rtol=0)

def test_multiple_rx_gradient_hermitian(self, tol, dev):
"""Tests that the gradient of multiple RX gates in a circuit yields the correct result
Expand All @@ -312,12 +305,13 @@ def test_multiple_rx_gradient_hermitian(self, tol, dev):
for idx in range(3):
qml.expval(qml.Hermitian([[1, 0], [0, -1]], wires=[idx]))

tape.trainable_params = {0, 1, 2}
# circuit jacobians
dev_jacobian = dev.adjoint_jacobian(tape)
expected_jacobian = -np.diag(np.sin(params))
trainable_params = {0, 1, 2}

# circuit Jacobians
calculated_val = self.process_and_compute_derivatives(dev, tape, trainable_params)
expected_val = -np.diag(np.sin(params))

assert np.allclose(dev_jacobian, expected_jacobian, atol=tol, rtol=0)
assert np.allclose(calculated_val, expected_val, atol=tol, rtol=0)

qubit_ops = [getattr(qml, name) for name in qml.ops._qubit__ops__]
ops = {qml.RX, qml.RY, qml.RZ, qml.PhaseShift, qml.CRX, qml.CRY, qml.CRZ, qml.Rot}
Expand All @@ -339,13 +333,15 @@ def test_multiple_rx_gradient_expval_hermitian(self, tol, dev):
)
)

tape.trainable_params = {0, 1, 2}
dev_jacobian = dev.adjoint_jacobian(tape)
expected_jacobian = np.array(
trainable_params = {0, 1, 2}

# circuit Jacobians
calculated_val = self.process_and_compute_derivatives(dev, tape, trainable_params)
expected_val = np.array(
[-np.sin(params[0]) * np.cos(params[2]), 0, -np.cos(params[0]) * np.sin(params[2])]
)

assert np.allclose(dev_jacobian, expected_jacobian, atol=tol, rtol=0)
assert np.allclose(calculated_val, expected_val, atol=tol, rtol=0)

qubit_ops = [getattr(qml, name) for name in qml.ops._qubit__ops__]
ops = {qml.RX, qml.RY, qml.RZ, qml.PhaseShift, qml.CRX, qml.CRY, qml.CRZ, qml.Rot}
Expand Down Expand Up @@ -376,9 +372,11 @@ def test_multiple_rx_gradient_expval_hamiltonian(self, tol, dev):

qml.expval(ham)

tape.trainable_params = {0, 1, 2}
dev_jacobian = dev.adjoint_jacobian(tape)
expected_jacobian = (
trainable_params = {0, 1, 2}

# circuit Jacobians
calculated_val = self.process_and_compute_derivatives(dev, tape, trainable_params)
expected_val = (
0.3 * np.array([-np.sin(params[0]), 0, 0])
+ 0.3 * np.array([0, -np.sin(params[1]), 0])
+ 0.4
Expand All @@ -387,7 +385,7 @@ def test_multiple_rx_gradient_expval_hamiltonian(self, tol, dev):
)
)

assert np.allclose(dev_jacobian, expected_jacobian, atol=tol, rtol=0)
assert np.allclose(calculated_val, expected_val, atol=tol, rtol=0)

qubit_ops = [getattr(qml, name) for name in qml.ops._qubit__ops__]
ops = {qml.RX, qml.RY, qml.RZ, qml.PhaseShift, qml.CRX, qml.CRY, qml.CRZ, qml.Rot}
Expand Down Expand Up @@ -527,94 +525,94 @@ def test_gradient_gate_with_multiple_parameters_hermitian(self, dev):
# the different methods agree
assert np.allclose(calculated_val, reference_val, atol=tol, rtol=0)

# @pytest.mark.skipif(not ld._CPP_BINARY_AVAILABLE, reason="Lightning binary required")
# def test_gradient_gate_with_multiple_parameters_hamiltonian(self, dev):
# """Tests that gates with multiple free parameters yield correct gradients."""
# x, y, z = [0.5, 0.3, -0.7]
@pytest.mark.skipif(not ld._CPP_BINARY_AVAILABLE, reason="Lightning binary required")
def test_gradient_gate_with_multiple_parameters_hamiltonian(self, dev):
"""Tests that gates with multiple free parameters yield correct gradients."""
x, y, z = [0.5, 0.3, -0.7]

# ham = qml.Hamiltonian(
# [1.0, 0.3, 0.3], [qml.PauliX(0) @ qml.PauliX(1), qml.PauliZ(0), qml.PauliZ(1)]
# )
ham = qml.Hamiltonian(
[1.0, 0.3, 0.3], [qml.PauliX(0) @ qml.PauliX(1), qml.PauliZ(0), qml.PauliZ(1)]
)

# tape = qml.tape.QuantumScript(
# [qml.RX(0.4, wires=[0]), qml.Rot(x, y, z, wires=[0]), qml.RY(-0.2, wires=[0])],
# [qml.expval(ham)],
# )
tape = qml.tape.QuantumScript(
[qml.RX(0.4, wires=[0]), qml.Rot(x, y, z, wires=[0]), qml.RY(-0.2, wires=[0])],
[qml.expval(ham)],
)

# trainable_params = {1, 2, 3}
trainable_params = {1, 2, 3}

# calculated_val = self.process_and_compute_derivatives(dev, tape, trainable_params)
# reference_val = self.calculate_reference(tape, trainable_params)
calculated_val = self.process_and_compute_derivatives(dev, tape, trainable_params)
reference_val = self.calculate_reference(tape, trainable_params)

# tol = 1e-3 if dev.C_DTYPE == np.complex64 else 1e-7
tol = 1e-3 if dev.C_DTYPE == np.complex64 else 1e-7

# # gradient has the correct shape and every element is nonzero
# assert len(calculated_val) == 3
# assert all(isinstance(v, np.ndarray) for v in calculated_val)
# assert np.count_nonzero(calculated_val) == 3
# gradient has the correct shape and every element is nonzero
assert len(calculated_val) == 3
assert all(isinstance(v, np.ndarray) for v in calculated_val)
assert np.count_nonzero(calculated_val) == 3

# # the different methods agree
# assert np.allclose(calculated_val, reference_val, atol=tol, rtol=0)
# the different methods agree
assert np.allclose(calculated_val, reference_val, atol=tol, rtol=0)

# def test_use_device_state(self, tol, dev):
# """Tests that when using the device state, the correct answer is still returned."""
def test_use_device_state(self, tol, dev):
"""Tests that when using the device state, the correct answer is still returned."""

# x, y, z = [0.5, 0.3, -0.7]
x, y, z = [0.5, 0.3, -0.7]

# with qml.tape.QuantumTape() as tape:
# qml.RX(0.4, wires=[0])
# qml.Rot(x, y, z, wires=[0])
# qml.RY(-0.2, wires=[0])
# qml.expval(qml.PauliZ(0))
with qml.tape.QuantumTape() as tape:
qml.RX(0.4, wires=[0])
qml.Rot(x, y, z, wires=[0])
qml.RY(-0.2, wires=[0])
qml.expval(qml.PauliZ(0))

# tape.trainable_params = {1, 2, 3}
tape.trainable_params = {1, 2, 3}

# dM1 = dev.adjoint_jacobian(tape)
dM1 = dev.adjoint_jacobian(tape)

# qml.execute([tape], dev, None)
# dM2 = dev.adjoint_jacobian(tape, use_device_state=True)
qml.execute([tape], dev, None)
dM2 = dev.adjoint_jacobian(tape, use_device_state=True)

# assert np.allclose(dM1, dM2, atol=tol, rtol=0)
assert np.allclose(dM1, dM2, atol=tol, rtol=0)

# def test_provide_starting_state(self, tol, dev):
# """Tests provides correct answer when provided starting state."""
# x, y, z = [0.5, 0.3, -0.7]
def test_provide_starting_state(self, tol, dev):
"""Tests provides correct answer when provided starting state."""
x, y, z = [0.5, 0.3, -0.7]

# with qml.tape.QuantumTape() as tape:
# qml.RX(0.4, wires=[0])
# qml.Rot(x, y, z, wires=[0])
# qml.RY(-0.2, wires=[0])
# qml.expval(qml.PauliZ(0))
with qml.tape.QuantumTape() as tape:
qml.RX(0.4, wires=[0])
qml.Rot(x, y, z, wires=[0])
qml.RY(-0.2, wires=[0])
qml.expval(qml.PauliZ(0))

# tape.trainable_params = {1, 2, 3}
tape.trainable_params = {1, 2, 3}

# dM1 = dev.adjoint_jacobian(tape)
dM1 = dev.adjoint_jacobian(tape)

# dev._pre_rotated_state = dev.state_vector # necessary for lightning.kokkos
dev._pre_rotated_state = dev.state_vector # necessary for lightning.kokkos

# qml.execute([tape], dev, None)
# dM2 = dev.adjoint_jacobian(tape, starting_state=dev._pre_rotated_state)
qml.execute([tape], dev, None)
dM2 = dev.adjoint_jacobian(tape, starting_state=dev._pre_rotated_state)

# assert np.allclose(dM1, dM2, atol=tol, rtol=0)
assert np.allclose(dM1, dM2, atol=tol, rtol=0)

# @pytest.mark.skipif(not ld._CPP_BINARY_AVAILABLE, reason="Lightning binary required")
# def test_provide_wrong_starting_state(self, dev):
# """Tests raise an exception when provided starting state mismatches."""
# x, y, z = [0.5, 0.3, -0.7]
@pytest.mark.skipif(not ld._CPP_BINARY_AVAILABLE, reason="Lightning binary required")
def test_provide_wrong_starting_state(self, dev):
"""Tests raise an exception when provided starting state mismatches."""
x, y, z = [0.5, 0.3, -0.7]

# with qml.tape.QuantumTape() as tape:
# qml.RX(0.4, wires=[0])
# qml.Rot(x, y, z, wires=[0])
# qml.RY(-0.2, wires=[0])
# qml.expval(qml.PauliZ(0))
with qml.tape.QuantumTape() as tape:
qml.RX(0.4, wires=[0])
qml.Rot(x, y, z, wires=[0])
qml.RY(-0.2, wires=[0])
qml.expval(qml.PauliZ(0))

# tape.trainable_params = {1, 2, 3}
tape.trainable_params = {1, 2, 3}

# with pytest.raises(
# qml.QuantumFunctionError,
# match="The number of qubits of starting_state must be the same as",
# ):
# dev.adjoint_jacobian(tape, starting_state=np.ones(7))
with pytest.raises(
qml.QuantumFunctionError,
match="The number of qubits of starting_state must be the same as",
):
dev.adjoint_jacobian(tape, starting_state=np.ones(7))

@pytest.mark.skipif(
device_name == "lightning.kokkos",
Expand Down Expand Up @@ -642,24 +640,24 @@ class TestAdjointJacobianQNode:
def dev(self, request):
return qml.device(device_name, wires=2, c_dtype=request.param)

# def test_finite_shots_warning(self):
# """Tests that a warning is raised when computing the adjoint diff on a device with finite shots"""
def test_finite_shots_warning(self):
"""Tests that a warning is raised when computing the adjoint diff on a device with finite shots"""

# dev = qml.device(device_name, wires=1, shots=1)
dev = qml.device(device_name, wires=1, shots=1)

# with pytest.warns(
# UserWarning, match="Requested adjoint differentiation to be computed with finite shots."
# ):
with pytest.warns(
UserWarning, match="Requested adjoint differentiation to be computed with finite shots."
):

# @qml.qnode(dev, diff_method="adjoint")
# def circ(x):
# qml.RX(x, wires=0)
# return qml.expval(qml.PauliZ(0))
@qml.qnode(dev, diff_method="adjoint")
def circ(x):
qml.RX(x, wires=0)
return qml.expval(qml.PauliZ(0))

# with pytest.warns(
# UserWarning, match="Requested adjoint differentiation to be computed with finite shots."
# ):
# qml.grad(circ)(0.1)
with pytest.warns(
UserWarning, match="Requested adjoint differentiation to be computed with finite shots."
):
qml.grad(circ)(0.1)

@pytest.mark.skipif(not ld._CPP_BINARY_AVAILABLE, reason="Lightning binary required")
def test_qnode(self, mocker, dev):
Expand Down Expand Up @@ -889,13 +887,13 @@ def circuit_ansatz(params, wires):
qml.QubitStateVector(unitary_group.rvs(2**4, random_state=0)[0], wires=wires)
qml.RX(params[0], wires=wires[0])
qml.RY(params[1], wires=wires[1])
qml.adjoint(qml.RX(params[2], wires=wires[2]))
qml.adjoint(qml.RX(params[2], wires=wires[2]), lazy=False)
qml.RZ(params[0], wires=wires[3])
qml.CRX(params[3], wires=[wires[3], wires[0]])
qml.PhaseShift(params[4], wires=wires[2])
qml.CRY(params[5], wires=[wires[2], wires[1]])
qml.adjoint(qml.CRZ(params[5], wires=[wires[0], wires[3]]))
qml.adjoint(qml.PhaseShift(params[6], wires=wires[0]))
qml.adjoint(qml.CRZ(params[5], wires=[wires[0], wires[3]]), lazy=False)
qml.adjoint(qml.PhaseShift(params[6], wires=wires[0]), lazy=False)
qml.Rot(params[6], params[7], params[8], wires=wires[0])
qml.adjoint(qml.Rot(params[8], params[8], params[9], wires=wires[1]))
qml.MultiRZ(params[11], wires=[wires[0], wires[1]])
Expand All @@ -908,7 +906,7 @@ def circuit_ansatz(params, wires):
qml.U1(params[15], wires=wires[0])
qml.U2(params[16], params[17], wires=wires[0])
qml.U3(params[18], params[19], params[20], wires=wires[1])
qml.adjoint(qml.CRot(params[21], params[22], params[23], wires=[wires[1], wires[2]]))
qml.adjoint(qml.CRot(params[21], params[22], params[23], wires=[wires[1], wires[2]]), lazy=False)
qml.SingleExcitation(params[24], wires=[wires[2], wires[0]])
qml.DoubleExcitation(params[25], wires=[wires[2], wires[0], wires[1], wires[3]])
qml.SingleExcitationPlus(params[26], wires=[wires[0], wires[2]])
Expand Down

0 comments on commit 2ed8493

Please sign in to comment.