Experiments in Visualizing Neural Network Architectures

I’m a visual learner, and I wish machine learning researchers put effort into showing off their work.

Here’s the original (top) and mine (bottom). I aim to communicate the specific mathematics in a visual interpretation of linear algebra.

The original paper’s architecture diagram for comparison. I think this diagram has less info and is harder to parse (particularly to double arrows going to the attention weights and code vector, when it’s actually easily expressed in a step-by-step way that’s easier to visually understand.