Advertisement
Felanpro

NeuralNetworkFrameworkBetaVersion1

Dec 9th, 2022 (edited)
952
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
Python 4.04 KB | Source Code | 0 0
  1. '''
  2. This neural network framework only nudges weights when it performs backpropogation. Also it only works on 1 single network output.
  3. '''
  4.  
  5. import numpy as np
  6. import random
  7.  
  8. class Layer:
  9.     def __init__(self, inputNodes, outputNodes):
  10.         self.weights = 0.1 * np.random.randn(inputNodes, outputNodes)
  11.         #self.weights = 2 + np.zeros((inputNodes, outputNodes))
  12.         self.biases = np.zeros((1, outputNodes))
  13.    
  14.     def forward(self, inputs):
  15.         self.output = np.dot(inputs, self.weights) + self.biases
  16.  
  17. class Activation_ReLU:
  18.     def forward(self, inputs):
  19.         self.output = np.maximum(0, inputs)    
  20.        
  21. learningRate = 0.0001
  22. def backwards(network, input_, desired):
  23.     currentLayer = len(network) - 1
  24.     dError = 2*(network[currentLayer].output - desired)
  25.    
  26.     gradients = np.zeros((len(network), 5))
  27.     gradients[currentLayer][0] = dError    
  28.                      
  29.     currentLayer = len(network) - 1
  30.     while currentLayer >= 0: # Per layer
  31.         print("Current layer: ", currentLayer + 1, "--------------------------------------")
  32.         if type(network[currentLayer - 1]) == Activation_ReLU:
  33.             pass
  34.         else:
  35.             if currentLayer != 0:
  36.                
  37.                 #Nudge the weights
  38.                 for neuronCurrentLayer in range(len(network[currentLayer].output[0])): # Per neuron in current layer
  39.                     #print("Neuron ", neuronCurrentLayer + 1, ": ")
  40.                     for neuronPreviousLayer in range(len(network[currentLayer - 1].output[0])): # Per neuron in previous layer
  41.                         network[currentLayer].weights[neuronPreviousLayer][neuronCurrentLayer] -= network[currentLayer - 1].output[0][neuronPreviousLayer] * gradients[currentLayer][neuronCurrentLayer] * learningRate
  42.                         #print(network[currentLayer].weights[neuronPreviousLayer][neuronCurrentLayer])    
  43.                
  44.                
  45.                 # Calculate gradients for every neuron in the next layer you're going to adjust
  46.                 for neuronCurrentLayer in range(len(network[currentLayer].output[0])): # Per neuron in current layer
  47.                     #print("Neuron ", neuronCurrentLayer + 1, ": ")
  48.                     for neuronPreviousLayer in range(len(network[currentLayer - 1].output[0])): # Per neuron in previous layer
  49.                         gradients[currentLayer - 1][neuronPreviousLayer] += network[currentLayer].weights[neuronPreviousLayer][neuronCurrentLayer] * gradients[currentLayer][neuronCurrentLayer]  
  50.                    
  51.             '''
  52.            else: #Special for first layer
  53.                for outputNodessss in range(len(network[0].output[0])):
  54.                    print("Neuron ", outputNodessss + 1, ": ")
  55.                    for inputNodessss in range(len(input_)):
  56.                        print(network[0].weights[inputNodessss][outputNodessss])
  57.            '''
  58.        
  59.         '''
  60.        print("Weights: ", network[currentLayer].weights)
  61.        print("Layer output: ", network[currentLayer].output[0])
  62.        print("Gradients: ", gradients[currentLayer])
  63.        '''
  64.        
  65.         currentLayer -= 1 #Go to previous layer
  66.    
  67.     '''
  68.    print("-----------------------------------")
  69.    print("Gradients total: ")
  70.    print(gradients)
  71.    '''
  72.     print("Error: ", (network[len(network) - 1].output[0] - desired))
  73.        
  74. #Create neural network
  75. inputs = [4, 6, 1, 3, 9, 2, 3, 7, 10, 34]
  76. desired = [8, 12, 2, 6, 18, 4, 6, 14, 20, 68]
  77. layer1 = Layer(1, 3)
  78. activation1 = Activation_ReLU()
  79. layer2 = Layer(3, 3)
  80. activation2 = Activation_ReLU()
  81. layer3 = Layer(3, 1)
  82.  
  83.  
  84. #Train the network
  85. for x in range(500):
  86.     for iteration in range(10):
  87.         layer1.forward(inputs[iteration])
  88.         layer2.forward(layer1.output)
  89.         layer3.forward(layer2.output)
  90.         backwards([layer1, layer2, layer3], inputs[iteration], desired[iteration])
  91.        
  92. #Test the network
  93. userInput = 49
  94. layer1.forward(userInput)
  95. layer2.forward(layer1.output)
  96. layer3.forward(layer2.output)
  97. print("Guess: ", layer3.output)
  98.  
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement