Pretty sure Llama is all Meta. But tons of people have dropped forks and fine tunes of it since it’s FOSS. This is likely NVIDIA just working off of their base model and applying their RLHF tricks or whatever to make it better.