Math

A perpetual pain point for people working on computer vision is dealing with affine transformations. This post provides a simple example of how to reason about them.

In-depth explanation of the math behind the RWKV model, with PyTorch implementations, plus a discussion of numerical stability.

Simple overview of what a PID controller is, how it works, and how to make one yourself.

A walkthrough of how to optimize the log-sum-exp function in PyTorch.

A comparison of Hidden Markov Models and Conditional Random Fields, two kinds of probabilistic graphical models.

A demo of how to code the Viterbi algorithm in Numpy.