Experimentation of CNN with Fashion MNIST Dataset

Prequisites

Tensorflow 1.12
Keras
Python 3.x

Dataset

Fashion-MNIST is dataset of Zalando's article images—consisting of a training set of 60,000 examples and a test set of 10,000 examples. Each example is a 28x28 grayscale image, associated with a label from 10 classes. Zalando intends Fashion-MNIST to serve as a direct drop-in replacement for the original MNIST dataset for benchmarking machine learning algorithms. It shares the same image size and structure of training and testing splits.

How to Use

Open this repository notebook
Run each step by pressing "play" button or CTRL + ENTER
Change the parameters in "Parameters Section"

Run the Models

Go to section running models
Select model that want to evaluate
Run evaluate

Training Steps

Importing required modules

import tensorflow as tf
from tensorflow import keras
from sklearn.model_selection import StratifiedShuffleSplit
from keras import backend as k

import numpy as np
import matplotlib.pyplot as plt

Load Datasets

(imageTrain, labelTrain), (imageTest, labelTest) = tf.keras.datasets.fashion_mnist.load_data()

Make array of labels and split training data into validation data

labels = {0 : "T-shirt/top", 1: "Trouser", 2: "Pullover", 3: "Dress", 4: "Coat", 5: "Sandal", 6: "Shirt", 7: "Sneaker", 8: "Bag", 9: "Ankle Boot"}

imageTrain = np.expand_dims(imageTrain, -1)
imageTest = np.expand_dims(imageTest, -1)

sss = StratifiedShuffleSplit(n_splits=5, random_state=0, test_size=1/6)
trainIndex, validIndex = next(sss.split(imageTrain, labelTrain))
imageValid, labelValid = imageTrain[validIndex], labelTrain[validIndex]
imageTrain, labelTrain = imageTrain[trainIndex], labelTrain[trainIndex]

Perform Mean Subtraction and Normalization of Datasets (a)

Scaling down range of input value into between 0 and 1 instead of 0 - 255

imageTrain = imageTrain / 255
imageValid = imageValid / 255
imageTest = imageTest / 255

Mean Substraction of data

meanSubt = np.mean(imageTrain)

imageTrain = imageTrain - meanSubt
imageValid = imageValid - meanSubt
imageTest = imageTest - meanSubt

Normalization

stdDev = np.std(imageTrain)

imageTrain = imageTrain / stdDev
imageValid = imageValid / stdDev
imageTest = imageTest / stdDev

Mean subtraction and normalization using image training data to make sure it's same across all datasets that we use. This process necessary because we want to make our data sparse before we train.

Choosing Initializer (b)

We experiment with two weight initialization: He initialization and Xavier Initialization

We can see the result in below table:

	Xavier Initializer	He Initializer
Accuracy	0.9663	0.9675
Loss	0.0891	0.0863
Validation Accuracy	0.9194	0.9184
Validation Loss	0.284	0.2929
Test Accuracy	0.9138	0.9159
Test Loss	0.3143	0.3164

Experimentation with network configuration (c)

	32 Filter + 2 Layers	64 Filter + 2 Layers
Accuracy	0.9663	0.9675
Loss	0.0891	0.0863
Validation Accuracy	0.9194	0.9184
Validation Loss	0.284	0.2929
Test Accuracy	0.9138	0.9159
Test Loss	0.3143	0.3164

ivokun / fashion-mnist-tensorflow Goto Github PK

fashion-mnist-tensorflow's Introduction

Experimentation of CNN with Fashion MNIST Dataset

Prequisites

Dataset

How to Use

Run the Models

Training Steps

Importing required modules

Load Datasets

Make array of labels and split training data into validation data

Perform Mean Subtraction and Normalization of Datasets (a)

Choosing Initializer (b)

Experimentation with network configuration (c)

Experimentation with gradient optimization techniques (d)

Experimentation with activation functions (e)

Experimentation with regularization techniques (f)

fashion-mnist-tensorflow's People

Contributors

Watchers

Recommend Projects

Recommend Topics

Recommend Org