Learn Challenge: Creating a Perceptron | Neural Network from Scratch

Since the goal is to implement a multilayer perceptron, defining a Perceptron class helps organize and initialize the model efficiently. The class will contain a single attribute, layers, which is a list of Layer objects representing the structure of the network:

class Perceptron:
    def __init__(self, layers):
        self.layers = layers

The variables used to initialize the layers are:

input_size: the number of input features;
hidden_size: the number of neurons in each hidden layer (both hidden layers will have the same number of neurons in this case);
output_size: the number of neurons in the output layer.

The structure of the resulting multilayer perceptron will include:

Input layer → receives the data;
Two hidden layers → process the inputs and extract patterns;
Output layer → produces the final prediction.

Task

Swipe to start coding

Your goal is to set up the basic structure of a multilayer perceptron (MLP) by implementing the code for its layers.

Follow these steps carefully:

Initialize layer parameters inside the __init__() method:
- Create the weight matrix with shape (n_neurons, n_inputs);
- Create the bias vector with shape (n_neurons, 1);
- Fill both with random values from a uniform distribution in the range $[-1, 1)$ using np.random.uniform().
Implement forward propagation inside the forward() method:
- Compute the raw output of each neuron using the dot product:
```
np.dot(self.weights, self.inputs) + self.biases
```
- Apply the assigned activation function to this result and return the activated output.
Define the perceptron layers:
- Create two hidden layers, each containing hidden_size neurons and using the ReLU activation function;
- Create one output layer with output_size neuron(s) and the sigmoid activation function.

Solution

Everything was clear?

Thanks for your feedback!

Section 2. Chapter 4

single

Ask AI

Ask anything or try one of the suggested questions to begin our chat

Awesome!

Completion rate improved to 4

Swipe to show menu

class Perceptron:
    def __init__(self, layers):
        self.layers = layers