Repository containing various small projects implemented using PyTorch related to: Deep Learning in Computer Vision. Each project contains a report that gathers the needed theoritical background and the different implementation's steps. Code is run on a GeForce GTX TITAN X 12GB GPU. The results are sometimes relatively bad but the aim here is to understand the main concepts.
- Exercises to get familiar with Deep Learning and PyTorch, that can be found under
exercises/ - Hotdog / nothotdog classifier
- Digits detection using sliding windows
- Nuclei segmentation inspired from the 2018 Data Science Bowl
- CycleGAN that converts horses to zebras and vice versa