Building GPT from Scratch — Course Lab
Course lab implementing a GPT-style transformer decoder, including causal self-attention, from scratch.
Course: Procesamiento de Lenguaje Natural, Departamento de Ciencia de la Computación, Pontificia Universidad Católica de Chile Professor: Marcelo Mendoza
Course-provided lab material, archived here for reference — full credit for authorship goes to the professor and course staff.