Udacity - Machine learning Nano Degree Program : Project-5

This is fifth project in the series of the projects listed in Udacity- Machine Learning Nano Degree Program.

In this project, I have classified images from the CIFAR-10 dataset. The dataset consists of airplanes, dogs, cats, and other objects. The dataset was first preprocessed, then trained a convolutional neural network on all the samples. I have normalized the images, one-hot encoded the labels, build a convolutional layer, max pool layer, and fully connected layer.

In this project i have applied deep learning techniques to train a model to classify images using convolutional neural network algorithm. I have also learnt how to apply my knowledge of neural networks on real datasets using TensorFlow, an open source Deep Learning library created by Google.

Achievements:

• Built an image classification bot using convolutional neural networks to classify images from the CIFAR-10 dataset.
• Achieved accuracy of 60.42 %.

Things i have learnt by completing this project:

• How to apply deep learning techniques: Convolutional Neural Network algorithms.
• How to use tensorflow library.
• How to preprocess given data.
• How to analyze model's performance.
• How to optimise Convolutional Neural Network algortithm, to ensure increase in postive results.

• Project 0 : Titanic Survivals Prediction
• Project 1 : Boston's Houses Prediction
• Project 2 : Charity Donors Prediction
• Project 3 : Creating Customer Segments
• Project 4 : Smart Cab
• Project 6 : Stock Price Predictor

This project uses the following software and Python libraries:

• Python 2.7
• NumPy
• pandas
• Tensor-flow
• Jupyter Notebook