Spring 2018 ULab Computer Vision

Please fork this repository.

Running on your own computer

1. Fork this github repository (make sure git is setup!)
2. git clone [repo url]
3. git remote add ulab https://github.com/berkeley-ulab/sp18-ulab-computervision.git
4. Setup Conda via (this link)[https://conda.io/miniconda.html]
5. cd setup and execute setup.sh.
6. Run source activate ulab.
7. Run jupyter lab in this repository.

Updating Your Forked Repository

1. git pull ulab master

Navigating Jupyter Notebooks

This link has a quick start on Keyboard navigation shortcuts.

Curriculum Agenda

• Numpy
  • Manipulating matrices and vectors
  • Indexing
  • Broadcasting
• Matplotlib
  • Plotting images
  • Line/scatter plots
• PyTorch
  • Creating and manipulating Tensors - very similar to Numpy
  • nn.Linear
  • Variable and autograd
  • Learn how to train a linear model and visualize results
• Machine Learning (basics)
  • Data - how data is stored and accessed - with particular focus on images
  • Model
  • Training/Testing
  • Train a model on MNIST

Order of notebooks

0. NeuralStyleTransfer (demo)
1. numpy_matplotlib
2. pytorch
3. autograd_tutorial
4. introml