Single Layer Perceptron - logic AND (train)

train_slp.py

for i in range(NUM_ITER):
    y_pred = np.dot(x, W) + b
        
    #apply activation
    y_pred[y_pred > 0] = 1
    y_pred[y_pred <= 0] = 0
    
    #calculate error
    err = y - y_pred
    
    #stop if error = 0
    if np.sum(err) == 0:
        break
    
    #update w & b
    delta_W = learning_rate * np.dot(np.transpose(x) , err)
    delta_b = learning_rate * np.sum(err)
    W = W + delta_W
    b = b + delta_b
    
    print ("Iterasi ke-" + str(i), err, W, b)