Live Online Classes in Design, Coding & AI — Small Classes, Free Retakes
Pass the PMP® Exam on Your First Try — Expert-Led Training
Overview
Google, IBM & Meta Certificates — All 10,000+ Courses at 40% Off
One annual plan covers every course and certificate on Coursera. 40% off for a limited time.
Get Full Access
Learn about the Quasi-Hyperbolic Adam (QHAdam) optimizer in this 54-minute tutorial that breaks down both theoretical concepts and practical implementation. Explore the mathematical foundations, performance characteristics, and motivations behind QHAdam and QHM (Quasi-Hyperbolic Momentum) optimizers through detailed formula breakdowns and code examples. Follow along with both naive and official PyTorch implementations from the authors, examining the optimizer's architecture and real-world applications. Gain hands-on experience working with code examples available on GitHub while understanding the research paper's key findings and practical implications for machine learning optimization tasks.
Syllabus
# Table of Content- Introduction: 0:00- Overview of QHM and QHAdam: 1:12- QHM and QHAdam Performance: 3:07- QHM and QHAdam Formula Breakdown: 5:34- Naive Implementation of QHM and QHAdam: 16:26- Official QHM Pytorch Implementation from Authors: 23:17- Official QHAdam Pytorch Implementation from Authors: 29:57- Motivation Behind QHM: 39:08- Conclusion: 53:02- Spooky Stuff :
Taught by
Yacine Mahdid