The aim is to predict type of clothes on the image
Fashion dataset consists of 70,000 images divided into 60,000 training and 10,000 testing samples. Dataset sample consists of 28x28 grayscale image, associated with a label.
There are 10 classes (labels) available:
- 0 => T-shirt/top
- 1 => Trouser
- 2 => Pullover
- 3 => Dress
- 4 => Coat
- 5 => Sandal
- 6 => Shirt
- 7 => Sneaker
- 8 => Bag
- 9 => Ankle boot
Each image is 28 pixels in height and 28 pixels in width, for a total of 784 pixels in total. Each pixel has a single pixel-value associated with it, indicating the lightness or darkness of that pixel, with higher numbers meaning darker. This pixel-value is an integer between 0 and 255.
- Data import
- Load libraries
- Load dataset
- Data visualizing
- Display dataframe
- Display random images
- Model training
- Prepare train, validation and test datasets
- Create neural network model
- Compile model
- Train model
- Model evaluating
- Test data
- Print heatmap
- Print classification report
- By using (32,3,3) kernel with dropout, we were able to achieve:
- Accuracy: 89.55%
- Test accuracy: 90.10%
This implementation was inspired by Kirill Eremenko, Hadelin de Ponteves, Dr. Ryan Ahmed, Ph.D., MBA, SuperDataScience Team, Rony Sulca Machine Learning Practical Udemy course