Why High-Level Training Frameworks are a Crutch (Part 1)
Hosted by Zach Mueller
What you'll learn
Helpful, at first
Having extremely high abstraction levels are great for beginners when it comes to training models, however
Debug hell
When you get an error, the inexperienced spend more time debugging than going
Run and hope for the best
Hit an error? Beginners might not know what that means or why
Why this topic matters
The transformers Trainer, axolotl & TRL's one-line CLI. These are all excellent ways to get off the ground with training Deep Learning models. However, if you're new to the landscape and spend too much time with these, quickly you'll face errors you don't understand nor know how to fix them. In this series I'll be discussing the aspects of high and low level APIs and their usefulness's.
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 blogs, courses, and given talks on distributed training and PyTorch throughout my career.
Go deeper with a course
Keep exploring
