Template.py

# This is the template file for the Flappy Bird project. You can use this file to implement your own solution.

import random
import torch
import torch.nn as nn

MUTATION_RATE = 0.4  # Probability of each weight/bias being mutated
MUTATION_STRENGTH = 0.1
N = 5  # Number of birds to select for next round

""" Each bird contains a neural network model that decides when to jump (variable name is 'model'). Each bird also contains a 'deathTime' that records how long the bird survived for. 
    This should be used to determine which birds are the best. """

# Neural network model for each bird
class Model(nn.Module):
    def __init__(self, in_channels, out_channels):
        """ Initializes the neural network model with correct input and output sizes. """
        super(Model, self).__init__()
        # TODO: Define the layers of the neural network here

    
    def forward(self, input):
        """
        Forward pass of the neural network.
        Input is [closest_pipe_x, closest_pipe_y, y_position, velocity]. All values are normalized between 0 and 1.
        Output is a single value between 0 and 1. If output < 0.5, the bird should jump.
        """
        # TODO: Define the forward pass of the neural network here
        return 0 # Placeholder return value
    

def mutate_bird(bird):
    """ 
    Mutates the bird's neural network model based on the mutation rate and strength (constants above).
    """
    new_bird = bird.clone()
    # TODO: Implement mutation of the bird's model here (loop through new_bird.model.parameters() and mutate each parameter). bird.clone() is used to copy the bird without modifying the original.
    return new_bird

def select_next_round(birds):
    """ 
    Selects the N best birds to be used for the next round.
    """
    best_birds = []
    # TODO: Sort the birds based on their deathTime and return the top performers OR use a probability distribution based on deathTime to select birds (so that even lower performing birds have a chance of being selected).
    best_birds = birds[:N]  # Placeholder selection (Remove this line)
    return best_birds