/dl_for_computer_vision

Revision resources for Engineering Part IIB/Module 4F12: Computer Vision

Primary LanguageJupyter NotebookMIT LicenseMIT

Deep Learning for Computer Vision

This repository is a companion to a pair of lectures given at the University of Cambridge, which can be viewed here:

1: Multi-layer Perceptrons

2: Convolutional Neural Nets

The lectures were given as part of Engineering Tripos Part IIB, 4F12: Computer Vision.

In this repository you will find the code to reproduce most of the visualizations and experiments shared in the lectures, as well as two Jupyter Notebooks providing interactive lecture notes covering a variety of topics:

  1. Perceptrons
  2. Classification
  3. Backpropagation
  4. Gradient Descent
  5. Optimisation
  6. Datasets (MNIST, CIFAR-10, ImageNet, COCO)
  7. Convolution
  8. ResNet

Getting Started

You should be able to run all the scripts and the notebooks by running these commands:

pip install wheel
pip install -r requirements.txt

And then following the instructions to install PyTorch. You can run the scripts on their own, but they are largely meant to be run as part of one of the two Jupyter notebooks. Learn more about Jupyter.