PyTorch Hooks from A to `full_backward_hook`

Hosted by Zach Mueller

111 students

Share this lesson

Go deeper with a course

From Scratch to Scale: Hands on Distributed Training from the ground up
Zachary Mueller
View syllabus

What you'll learn

Just what are PyTorch Hooks?

What are hooks in PyTorch? And have you already used them? (Surprisingly: yes)

Why are hooks useful?

Learn the most common ways hooks are currently used and why

Learn to implement them yourself

It's only useful if you know how to do it! We'll make sure by the end you can write your own hooks.

Why this topic matters

Hooks are important in modern Deep Learning, but many (as proven by an earlier X poll) have either: never written one, or even know *what they do*. Truthfully, if you are using anything in the modern DL space (big model inference, large-scale training) hooks make their way everywhere. In this lesson, we're going to dive into them and how they're being used around you.

You'll learn from

Zach Mueller

Instructor, Technical Lead at Hugging Face

I've been in the field for almost a decade now. I first started in the fast.ai community, quickly learning how modern-day training pipelines are built and operated. Then I moved to Hugging Face, where I'm the Technical Lead on the accelerate project and manage the transformers Trainer.


I've written numerous blogscourses, and given talks on distributed training and PyTorch throughout my career.

Previously at

Hugging Face
Accenture

Watch the recording for free

By continuing, you agree to Maven's Terms and Privacy Policy.

© 2025 Maven Learning, Inc.