This repo contains a low-rank adapter for LLaMA-7b fit on the Stanford Alpaca dataset translated into Japanese.
It doesn't contain the foundation model itself, so it's MIT licensed.
Instructions for running it can be found at https://github.com/kunishou/Japanese-Alpaca-LoRA.