Private beta

I tried the Llama V2 on the same day it was released. It is good to see such a move from Meta; however, there may be some negative consequences about releasing such capable models to the public, but this is another discussion. The new Llama’s 7B model is better than I expected. I tried it on Google Colab using this repo. The results are good for a small model. I am looking forward to seeing a detailed analysis.