Machine Learning Projects

Computer Vision - Object Detection

I collected suitable data and annotated it in order to train a YOLO model to detect a water bottle, orange hammer, or rock hammer.

View Project →

2D Image Autoencoder

During my research, I needed to learn how an autoencoder model works and how to effectively train one. I used a publicly available dataset of paintings to train the model and observe its progress during training.

View Project →

3D Sparse Voxel Autoencoder

Inspired by the machine learning research I did, I created a 3D autoencoder model using the spconv framework with PyTorch. The key innovation is training a sparse model, because most of a voxelized model is empty. This massively decreases VRAM usage and increases training speed.

View Project →