mindquantum.framework.MQN2Layer

class mindquantum.framework.MQN2Layer(expectation_with_grad, weight='normal')

MindQuantum trainable layer.

Quantum neural network include encoder and ansatz circuit. The encoder circuit encode classical data into quantum state, while the ansatz circuit act as trainable circuit. This layer will calculate the square of absolute value of expectation automatically.

Parameters

• expectation_with_grad (GradOpsWrapper) – a grad ops that receive encoder data and ansatz data and return the expectation value and gradient value of parameters respect to expectation.

• weight (Union[Tensor, str, Initializer, numbers.Number]) – Initializer for the convolution kernel. It can be a Tensor, a string, an Initializer or a number. When a string is specified, values from 'TruncatedNormal', 'Normal', 'Uniform', 'HeUniform' and 'XavierUniform' distributions as well as constant 'One' and 'Zero' distributions are possible. Alias 'xavier_uniform', 'he_uniform', 'ones' and 'zeros' are acceptable. Uppercase and lowercase are both acceptable. Refer to the values of Initializer for more details. Default: 'normal'.

Inputs:

• enc_data (Tensor) - Tensor of encoder data that you want to encode into quantum state.

Outputs:

Tensor, The square of absolute value of expectation value of the hamiltonian.

Raises

ValueError – If length of shape of weight is not equal to 1 and shape[0] of weight is not equal to weight_size.

Supported Platforms:

GPU, CPU

Examples

>>> import numpy as np
>>> import mindspore as ms
>>> from mindquantum.core.circuit import Circuit
>>> from mindquantum.core.operators import Hamiltonian, QubitOperator
>>> from mindquantum.framework import MQN2Layer
>>> from mindquantum.simulator import Simulator
>>> ms.set_seed(42)
>>> ms.set_context(mode=ms.PYNATIVE_MODE, device_target="CPU")
>>> enc = Circuit().ry('a', 0).as_encoder()
>>> ans = Circuit().h(0).rx('b', 0).as_ansatz()
>>> ham = Hamiltonian(QubitOperator('Z0'))
>>> sim = Simulator('mqvector', 1)
>>> grad_ops = sim.get_expectation_with_grad(ham, enc + ans)
>>> enc_data = ms.Tensor(np.array([[0.1]]))
>>> net =  MQN2Layer(grad_ops)
>>> opti = ms.nn.Adam(net.trainable_params(), learning_rate=0.1)
>>> train_net = ms.nn.TrainOneStepCell(net, opti)
>>> for i in range(100):
...     train_net(enc_data)
>>> net.weight.asnumpy()
array([1.5646162], dtype=float32)
>>> net(enc_data)
Tensor(shape=[1, 1], dtype=Float32, value=
[[ 3.80662982e-07]])