Last active
December 5, 2022 18:47
-
-
Save danzeeeman/3acdd8e3a977ee60eb2fa81f7e406a80 to your computer and use it in GitHub Desktop.
write a gradient descent neural network in numpy
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
# Import NumPy | |
import numpy as np | |
# Define the architecture of the network | |
n_inputs = 2 | |
n_hidden = 3 | |
n_outputs = 1 | |
# Initialize the weights and biases | |
weights = { | |
'hidden': np.random.randn(n_inputs, n_hidden), | |
'output': np.random.randn(n_hidden, n_outputs), | |
} | |
biases = { | |
'hidden': np.random.randn(n_hidden), | |
'output': np.random.randn(n_outputs), | |
} | |
# Define the activation function | |
def sigmoid(x): | |
return 1 / (1 + np.exp(-x)) | |
# Define the derivative of the activation function | |
def sigmoid_prime(x): | |
return sigmoid(x) * (1 - sigmoid(x)) | |
# Set the learning rate | |
learning_rate = 0.5 | |
# Implement the forward propagation step | |
def forward_propagate(inputs): | |
hidden_inputs = np.dot(inputs, weights['hidden']) + biases['hidden'] | |
hidden_outputs = sigmoid(hidden_inputs) | |
output_inputs = np.dot(hidden_outputs, weights['output']) + biases['output'] | |
outputs = sigmoid(output_inputs) | |
return outputs | |
# Implement the backpropagation step | |
def backpropagate(inputs, targets, outputs): | |
# Calculate the error in the output | |
output_errors = targets - outputs | |
# Calculate the error in the hidden layer | |
hidden_errors = np.dot(output_errors, weights['output'].T) * sigmoid_prime(hidden_outputs) | |
# Update the weights and biases | |
weights['output'] += learning_rate * np.dot(hidden_outputs.T, output_errors) | |
biases['output'] += learning_rate * output_errors | |
weights['hidden'] += learning_rate * np.dot(inputs.T, hidden_errors) | |
biases['hidden'] += learning_rate * hidden_errors | |
# Train the network using the gradient descent algorithm | |
for epoch in range(10000): | |
# Generate some random input data | |
inputs = np.random.randn(n_inputs) | |
targets = np |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment