Bamboo-Nano
This is an experiment to see if a smaller vocabulary affects model training, inspired by the "TinyStories" paper.
Training Data
The various datasets we trained on are contained within this readme.md file. These are all permissively licensed datasets.