/deep-learning-resources

Collection of resources for getting started with PyTorch

The UnlicenseUnlicense

deep-learning-intro

A collaborative collection of resources for learning the theory and practice of deep learning. Started by Audrey Beard, a PhD student of Computer Science at Rensselaer Polytechnic Institute in Troy, NY.

PyTorch

This repo focuses on PyTorch, since that's what we use in the lab. To get started with PyTorch, check out pytorch-intro.md

Deep Neural Network Development

There are myriad sources for learning about deep learning. Here's a running list of the best ones we can find. Some of these are hidden behind paywalls. You should go to the site to help out the creators, but if you need access and can't get it for some reason, I have copies of all these resources.

A Recipe for Training Neural Networks

This blog post by Andrej Karpathy of Stanford (and one author of the ImageNet paper) is a great place to start with if you're just learning or even experienced and stuck.

Troubleshooting Deep Neural Networks

This presentation by Josh Tobin is also a great resource if you're stuck (or before you get stuck!)

Things I wish we had known before we started our first Machine Learning project

This Medium article by Aseem Bansal

How to Unit Test Machine Learning Code

This Medium article by Chase Roberts

Practical Advice for Building Deep Neural Networks

This blog post by David Wingate and Matt Holt at BYU

Top 6 Errors Novice Machine Learning Engineers Make

This Medium article by Christopher Dossman

CS231n: Convolutional Neural Networks for Visual Recognition

This Stanford Course is taught by Fei-Fei Li

CS224n: Natural Language Processing with Deep Learning

This Stanford Course is taught by Chris Manning

CS285 Fall 2019

This YouTube playlist was published at UC Berkeley, where Jitendra Malik teaches

CS 330: Deep Multi-Task and Meta Learning

This Stanford course is taught by Chelsea Finn

Twitter thread about conda-forge

This Twitter thread addresses some issues that conda users might fact

Important Papers

  • ImageNet
    • Arguably pivotal in launching the present wave of computer vision research
    • Enormous and ever-growing image dataset
    • Yearly challenges for several tasks
    • Now a Kaggle competition
    • See image-net.org for more info