Why Adaptive Optimizers Exist — a Tangible lesson.
🏔
8
Why momentum/AdamW can outperform SGD — a Tangible lesson.
Tangible lessons
Interactive narrated lessons built with Tangible, the open-source framework for explorable lessons: https://github.com/scienceetonnante/tangible
Note Explore why momentum and AdamW outperform plain SGD in an interactive narrated lesson.