README.md

metadata

license: mit

This is the 15M parameter Llama 2 architecture model trained on the TinyStories dataset. These are converted from karpathy/tinyllamas. See the llama2.c project for more details.