Project description

On my personal time, both for fun and to improve my coding and deep learning skills, I have been playing around with some deep learning techniques. I have also participated in some Kaggle competitions to get acquainted with deep learning applications.
This page contains links to a few of my deep learning projects.

Technologies Used: Python, Tensorflow/Keras, Convolutional Neural Networks, Classical Machine Learning

Computer Vision Implementations

This section shows a few computer vision projects I have completed. In order to become more familiar with deep learning applied to computer vision, I wanted to create my own implementations of famous algorithms, as well as observe the impact of some hyperparameters (like learning rate, number of epochs, number of layers and units per layer, etc.). Note: the Github page for this project is currently in the works.

• General CNN with MNIST: Optimization of network parameters
• ResNet V1 with CIFAR-10: ResNet20, ResNet56, ResNet110
• ResNet V2 with CIFAR-10: ResNet20V2, ResNet56V2, ResNet110V2
• YOLO V1

Completed Kaggle competitions

My code and results can be found on my personal reporitory.

The projects I've completed:

• House Prices: Advanced Regression Techniques
• Digit Recognizer
• CIFAR-10 classifier

Other code

The following repositories contain code created as part of AI and Computer Vision classes at Polytechnique Montreal.

• INF8215 - Artificial Intelligence: Methods and algorithms (class)
• INF6804 - Computer Vision (class)