LLAMA

Introduction

LLaMA is a family of large language models (LLMs) developed by Meta AI. In their own words, Meta says -

Like other large language models, LLaMA works by taking a sequence of words as an input and predicts a next word to recursively generate text. Smaller models trained on more tokens — which are pieces of words — are easier to retrain and fine-tune for specific potential product use cases.

Note

1. We need to sign up and get access to LLAMA models by filling a form with Meta, even if we want to use it through Hugging face.
2. We are using LLAMA2-7b in AI3 for summarization. The implementation can be found in the ai3-hackathon repository under docs/nbs
3. More documentation can be found on the official site on Meta and also Hugging face.
4. The link to a 3rd party LLAMA playground is here: llama2.ai