This is a decoder-only transformer language model trained from scratch on the TinyStories dataset as part of the course assignment. The model implements the standard transformer architecture (decoder only), with several modern enhancements for efficient training and inference.
