Skip to content

Full Stack Deep Learning - Course Spring 2021

Info

This is the page for the 2021 edition of the course. For the 2022 edition, click here.

We've updated and improved our materials for our 2021 course taught at UC Berkeley and online.

Synchronous Online Course

We offered a paid synchronous option for those who wanted weekly assignments, capstone project, Slack discussion, and certificate of completion.

Enter your email below or follow us on Twitter to be the first to hear about future offerings of this option.

Week 1: Fundamentals

We do a blitz review of the fundamentals of deep learning, and introduce the codebase we will be working on in labs for the remainder of the class.

Reading:

How the backpropagation algorithm works

Week 2: CNNs

We cover CNNs and Computer Vision Applications, and introduce a CNN in lab.

Reading:

A brief introduction to Neural Style Transfer

Improving the way neural networks learn

Week 3: RNNs

We cover RNNs and applications in Natural Language Processing, and start doing sequence processing in lab.

Reading:

The Unreasonable Effectiveness of Recurrent Neural Networks

Attention Craving RNNS: Building Up To Transformer Networks

Week 4: Transformers

We talk about the successes of transfer learning and the Transformer architecture, and start using it in lab.

Reading:

Transformers from Scratch

Week 5: ML Projects

Our synchronous online course begins with the first "Full Stack" lecture: Setting up ML Projects.

Reading:

Rules of Machine Learning

ML Yearning (and subscribe to Andrew Ng's newsletter)

Those in the syncronous online course will have their first weekly assignment: Assignment 1, available on Gradescope.

Week 6: Infra & Tooling

We tour the landscape of infrastructure and tooling for deep learning.

Reading:

Machine Learning: The High-Interest Credit Card of Technical Debt

Those in the syncronous online course will have to work on Assignment 2.

Week 7: Troubleshooting

We talk about how to best troubleshoot training. In lab, we learn to manage experiments.

Reading:

Why is machine learning hard?

Those in the syncronous online course will have to work on Assignment 3.

Week 8: Data

We talk about Data Management, and label some data in lab.

Reading:

Emerging architectures for modern data infrastructure

Those in the syncronous online course will have to work on Assignment 4.

Week 9: Ethics

We discuss ethical considerations. In lab, we move from lines to paragraphs.

Those in the synchronous online course will have to submit their project proposals.

Week 10: Testing

We talk about Testing and Explainability, and set up Continuous Integration in lab.

Those in the synchronous online course will work on their projects.

Week 11: Deployment

We cover Deployment and Monitoring, and package up our model for deployment in lab.

Those in the synchronous online course will work on their projects.

Week 12: Research

We talk research, and set up robust monitoring for our model.

Those in the synchronous online course will work on their projects.

Week 13: Teams

We discuss ML roles and team structures, as well as big companies vs startups.

Week 14-16: Projects

Those in the synchronous online course will submit 5-minute videos of their projects and associated write-ups by May 15.

Check out the course projects showcase.

Other Resources

Fast.ai is a great free two-course sequence aimed at first getting hackers to train state-of-the-art models as quickly as possible, and only afterward delving into how things work under the hood. Highly recommended for anyone.

Dive Into Deep Learning is a great free textbook with Jupyter notebooks for every part of deep learning.

NYU’s Deep Learning course has excellent PyTorch breakdowns of everything important going on in deep learning.

Stanford’s ML Systems Design course has lectures that parallel those in this course.

The Batch by Andrew Ng is a great weekly update on progress in the deep learning world.

/r/MachineLearning/ is the best community for staying up to date with the latest developments.