Basic Perceptron Learning for AND Gate

ann.py

import numpy as np

class Perceptron:
    
    def __init__(self,Weights,Biases):
        
        self.Weights = Weights
        self.Biases = Biases
        
    def Train(self, Training, LearningRate):
        
        y_in = np.dot(Training[:,0:2], self.Weights) + self.Biases
        
        # This is meant for a simple two-input AND gate, hence only first two columns of the training vector.
        # But can be easily extended to any linearly separable dataset.
        
        op = -1*np.ones(y_in.shape, dtype = int)
        op[y_in > 0] = 1 # Activation
        t = op.__eq__(Training[:,2].reshape((4,1))) # There should be a better way of  
                                                    # comparing arrays
        
        while not t.all():
            
            for i in range(Training.shape[0]):
                self.Weights = self.Weights + LearningRate*Training[i,2]* \
                                Training[i,0:2].reshape((2,1))
                self.Biases = self.Biases + LearningRate*Training[i,2]
            
            y_in = np.dot(Training[:,0:2], self.Weights) + self.Biases
            op = -1*np.ones(y_in.shape, dtype = int)
            op[y_in > 0] = 1
            t = op.__eq__(Training[:,2].reshape((4,1)))
        
    def Test(self,Testing):
        
        y_in = np.dot(Testing,self.Weights) + self.Biases
        op = -1*np.ones(y_in.shape,dtype = int)
        op[y_in > 0] = 1
        return op