ngminhphuc / Deep-Learning-using-Tensorflow-Keras-CNN-with-Fashion-MNIST-dataset-

Deep Learning using Tensorflow, Keras & CNN with Fashion MNIST dataset

Deep Learning using Tensorflow, Keras & CNN with Fashion MNIST dataset

This github repository has three major jupyter notebooks namely.

Fashion MNIST - CNN Model Prediction using Tensorflow,Keras- Part 1

• The objective of this jupyter notebook is to define,compile and evaluate a single layer convolutional neural network model using keras as the frontend and tensorflow as backend.

Fashion MNIST - Embedding & Visualization with TensorBoard - Part 2

• The objective of this jupyter notebook is to perform embedding and data visualisation using tensorboard and perform multiple tuning of parameters of t-SNE and PCA algorithms to better visualise the data and get more insight of this fashion MNIST database.For this we used a sample of 2500 images from the 10000 test dataset.

PCA (Principle Component Analysis) Visualisation

t-SNE (t distributed Stochastic Neighbour Embedding) Visualisation

Fashion MNIST - 3 Layer Convolution Neural Network Model using Tensorflow,Keras- Part 3

• The objective of this jupyter notebook is to define,compile & evaluate a 3 layer convolutional neural network(CNN) model and Visualise the validation accuracy and validation loss