pirmber / transformer-from-first-principlesView on GitHub
A GPT-style Transformer built entirely from scratch with NumPy, featuring manual forward/backward propagation through attention, LayerNorm, softmax, residuals, GELU, and cross-entropy. Includes Adam, gradient checking, visualization, and autoregressive text generation—without autograd or deep-learning frameworks.
☆12Sep 17, 2026Updated last month

Alternatives and similar repositories for transformer-from-first-principles

Users that are interested in transformer-from-first-principles are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.

Sorting:

Are these results useful?