> To quickly start working with a few selected customers, we retrained this model from Llama 2 the minute we got access to our entire cluster — the pretraining finished on the day of Mistral 7B release.
That does not confirm that it is not Mistral Medium. It is written to give that impression, but if it weren’t Mistral Medium, I would expect them to explicitly say so.
I think it does, as Mistral usually do their own models, and also, they couldn't fully commercialise Mistral Medium if it was LLama 2 based.
What do you mean? Didn’t llama2’s tos say something about if you had something like 600m users at that current moment in time, you can’t use it? If you don’t, you’re good to go.
Good point. Likely Llama trained on similar corpora to what Mistral medium uses