当然可以!这里是一个简单的带有ReLU激活函数的神经网络的示例:

import numpy as np

class ReLU():
    def forward(self, input):
        return np.maximum(0, input)

    def backward(self, input, grad_output):
        relu_grad = input > 0
        return grad_output * relu_grad

class NeuralNetwork():
    def __init__(self, input_size, hidden_size, output_size):
        self.weights1 = np.random.randn(input_size, hidden_size)
        self.bias1 = np.random.randn(hidden_size)
        self.weights2 = np.random.randn(hidden_size, output_size)
        self.bias2 = np.random.randn(output_size)
        self.relu = ReLU()

    def forward(self, input):
        hidden = np.dot(input, self.weights1) + self.bias1
        hidden = self.relu.forward(hidden)
        output = np.dot(hidden, self.weights2) + self.bias2
        return output

    def backward(self, input, output, target):
        grad_output = 2.0 * (output - target)
        grad_weights2 = np.dot(self.hidden.T, grad_output)
        grad_bias2 = np.sum(grad_output, axis=0)
        grad_hidden = np.dot(grad_output, self.weights2.T)
        grad_hidden = self.relu.backward(hidden, grad_hidden)
        grad_weights1 = np.dot(input.T, grad_hidden)
        grad_bias1 = np.sum(grad_hidden, axis=0)
        return grad_weights1, grad_bias1, grad_weights2, grad_bias2

    def train(self, input, target, learning_rate):
        output = self.forward(input)
        grad_weights1, grad_bias1, grad_weights2, grad_bias2 = self.backward(input, output, target)
        self.weights1 -= learning_rate * grad_weights1
        self.bias1 -= learning_rate * grad_bias1
        self.weights2 -= learning_rate * grad_weights2
        self.bias2 -= learning_rate * grad_bias2

这个神经网络有一个输入层,一个隐藏层和一个输出层。其中,隐藏层使用ReLU激活函数。在训练时,我们使用均方误差作为损失函数,并使用反向传播算法来计算梯度并更新权重和偏置项。

你能把我写一个带有激活函数的神经网络吗

原文地址: https://www.cveoy.top/t/topic/Zx0 著作权归作者所有。请勿转载和采集!

免费AI点我,无需注册和登录