Я пытаюсь написать скрипт Python, который берет образцы выходных данных из целевой функции (в моем случае функции sin(x)) и подгоняет/аппроксимирует эту функцию с использованием градиентного спуска и вариационных квантовых схем. br />Моя проблема в том, что все должно быть очень хорошо написано, и каждая функция кода отлично выполняет свою работу, но я получаю эту обратную трассировку (код ниже после обратной трассировки):
Python\Python312\Lib\site-packages\pennylane\_grad.py:157: UserWarning: Attempted to differentiate a function with no trainable parameters. If this is unintended, please add trainable parameters via the 'requires_grad' attribute or 'argnum' keyword.
warnings.warn(
Python\Python312\Lib\site-packages\pennylane\workflow\qnode.py:494: UserWarning: Received gradient_kwarg wires, which is not included in the list of standard qnode gradient kwargs.
warnings.warn(
Traceback (most recent call last):
File "TuTest3.py", line 65, in run_vqc
params, cost = opt.step(cost_function, init_params, dev=dev)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "Python\Python312\Lib\site-packages\pennylane\optimize\gradient_descent.py", line 93, in step
g, _ = self.compute_grad(objective_fn, args, kwargs, grad_fn=grad_fn)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "Python\Python312\Lib\site-packages\pennylane\optimize\gradient_descent.py", line 122, in compute_grad
grad = g(*args, **kwargs)
^^^^^^^^^^^^^^^^^^
File "Python\Python312\Lib\site-packages\pennylane\_grad.py", line 162, in __call__
self._forward = self._fun(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^
File "TuTest3.py", line 38, in cost_function
y_pred = circuit(params)
^^^^^^^^^^^^^^^
File "Python\Python312\Lib\site-packages\pennylane\workflow\qnode.py", line 1092, in __call__
self.construct(args, kwargs)
File "Python\Python312\Lib\site-packages\pennylane\workflow\qnode.py", line 929, in construct
self._qfunc_output = self.func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^
TypeError: qml_circuit() missing 1 required positional argument: 'wires'
Optimization Error:
VQC optimization failed.
Я понятия не имею, в чем проблема, и буду рад, если кто-нибудь сможет помочь.
Это код:
import traceback
import pennylane as qml
import numpy as np
# Define the target function (modify as needed)
def target_function(x):
return np.sin(x)
# Generate sample data
x_data = np.linspace(0, 2 * np.pi, 100)
y_data = target_function(x_data)
# Quantum circuit definition
def qml_circuit(params, wires):
"""
This circuit applies Ry rotations to each qubit based on parameters.
"""
for i, param in enumerate(params):
qml.RY(param, wires)
# Add your desired quantum operations here (e.g., entangling gates)
# Measurement
return [qml.expval(qml.PauliZ(wire)) for wire in wires]
# Cost function (Mean Squared Error)
def cost_function(params, dev):
"""
Calculates mean squared error between predicted and target function values.
"""
# Execute the quantum circuit with current parameters
circuit = qml.QNode(qml_circuit, wires=dev.wires, device=dev)
y_pred = circuit(params)
# Calculate mean squared error
return np.mean((y_data - y_pred) ** 2)
# VQC Optimization
def run_vqc(n_qubits, n_iters):
"""
Performs VQC optimization for function approximation.
Args:
n_qubits: Number of qubits in the quantum circuit.
n_iters: Number of optimization iterations.
"""
try:
# Define device and number of qubits
dev = qml.device("default.qubit", wires=n_qubits)
# Initial parameter guess
init_params = np.random.rand(n_qubits)
# Select optimizer (Adam in this case)
opt = qml.AdamOptimizer(stepsize=0.01)
# Perform optimization
for n in range(n_iters):
params, cost = opt.step(cost_function, init_params, dev=dev)
init_params = params
# Print optimized parameters
print("Optimized Parameters:", params)
return params
except Exception as e:
print(f"Optimization Error: ")
traceback.print_exc()
return None
# Run VQC with default parameters (modify as needed)
n_qubits = 2
n_iters = 100
optimized_params = run_vqc(n_qubits, n_iters)
if optimized_params is not None:
# Evaluation (if optimization successful)
dev = qml.device("default.qubit", wires=n_qubits)
circuit = qml.QNode(qml_circuit, wires=dev.wires, device=dev)
new_x = np.linspace(0, np.pi * 2, 50) # Test data points
y_pred = target_function(new_x) # True target values
predicted_y = circuit(optimized_params)
# Print comparison between predicted and target values
print("Predicted vs. Target:")
for x, y_t, y_p in zip(new_x, y_pred, predicted_y):
print(f"x: {x:.2f}, Target: {y_t:.4f}, Predicted: {y_p:.4f}")
else:
print("VQC optimization failed.")
Подробнее здесь: https://stackoverflow.com/questions/785 ... ane-python
