LLama

LLaMA is a new open-source, high-performance large language model from Meta AI - FAIR.LLaMA is actually a collection of foundation language models ranging from 7B to 65B parameters.The models have been trained on trillions of tokens, and show that it is possible to train state-of-the-art models using publicly available datasets exclusively, without resorting to proprietary and inaccessible…