03 Batches, Layers & Objects

Batches¶

it allows to be compatible with parallel operations or calculations, reason why use GPU and not CPU
it helps with generalization, inputs are features and to pass such samples of data in batches

batches help to keep prediction away from under or overfitting overfitting: when all of the samples are shown together underfitting: when samples are passes very few at a time ultimately will hurt generalization output

common batches size to not have deal with above mentioned issues: 32 or 16, rarely 128

Adding Layer¶

import numpy as np

inputs = [[1, 2, 3, 2.5],
          [2, 5, -1, 2],
          [-1.5, 2.7, 3.3, -0.8]]

weights1 = [[0.2, 0.8, -0.5, 1.0],
           [0.5, -0.91, 0.26, -0.5],
           [-0.26, -0.27, 0.17, 0.87]]
biases1 = [2, 3, 0.5]

weights2 = [[0.1, -0.14, 0.5],
           [-0.5, 0.12, -0.33],
           [-0.44, 0.73, -0.13]]
biases2 = [-1, 2, -0.5]

layer1_outputs = np.dot(inputs, nop.array(weights1).T) + biases1

layer2_outputs = np.dot(layer1_outputs, nop.array(weights2).T) + biases2

print(layer2_outputs)
#output
[[0.5031   -1.04185   -2.03875]
 [0.2434   -2.7332    -5.7633]
 [-0.99314 1.41254    -0.35655]]

Object Layer¶

import numpy as np

#input data to neural network
X = [[1, 2, 3, 2.5],
    [2, 5, -1, 2],
    [-1.5, 2.7, 3.3, -0.8]]

# Normalize dataset: 
# 1. scaling input data between -1 and +1
# 2. intialize biases as zero, 
# except when if output comes to be zero
# 3. take weights as random value

np.random.seed(0) #guassian distribution bounded around 0

#Initialize a layer

class Layer_Dense:
    def __init__(self, n_inputs, n_neurons): #features, random
        self.weights = 0.10 * np.random.randn(n_inputs, n_neurons)
        self.biases = np.zeros((1, n_neurons))
    def forward(self, inputs):
        self.output = np.dot(inputs, self.weights) + self.biases

# Note w/ this method we will not require transpose as we are defining the valid shape 


# print(0.10*np.random.randn(4,3))

#[[ 0.17640523  0.04001572  0.0978738 ]
# [ 0.22408932  0.1867558  -0.09772779]
# [ 0.09500884 -0.01513572 -0.01032189]
# [ 0.04105985  0.01440436  0.14542735]]

# print(np.zeros((1, 3)))
#array([[0., 0., 0.]])

layer1 = Layer_Dense(4, 5) #features: 4, random1: 5
layer2 = Layer_Dense(5, 2) #has to be: random1, random2: 2


layer1.forward(X)
print(layer1.output)
layer2.forward(layer1.output)
print(layer2.output)

#outputs
#[[ 0.10758131  1.03983522  0.24462411  0.31821498  0.18851053]
#  [-0.08349796  0.70846411  0.00293357  0.44701525  0.36360538]
#  [-0.50763245  0.55688422  0.07987797 -0.34889573 0.04553042]]
# [[ 0.148296   -0.08397602]
#  [ 0.14100315 -0.01340469]
#  [ 0.20124979 -0.07290616]]