I built this models/methods from scratch mostly using numpy and sometimes pytorch to clearly understand how everything works. Though the implementations are often guided by existing resources such as books and courses, I painstakingly introduced clarity, readability and clean code.

I have another repository for more NLP oriented tasks here. But all the advanced modeling are in this repository.

Transformers

1. Self-Attention with Numpy
2. GPT with Pytorch, sklearn, and numpy on Wine Reviews

Coursera Deep Learning with Pytorch

Here are my solutions. They are often different from the instructor's implementation for clarity and updated libraries.

Mastering OpenCV

Here are my solutions. They are often different from the instructor's implementation for clarity and updated libraries.