Meta Releases LLaMA to Democratize Large Language Model Research
Meta introduces LLaMA, a foundational large language model available in sizes ranging from 7 billion to 65 billion parameters. The release aims to make advanced AI research more accessible to researchers who lack massive computing infrastructure.
Meta publicly releases LLaMA (Large Language Model Meta AI), a state-of-the-art foundational large language model designed to help researchers advance their work in artificial intelligence. By offering smaller, highly performant models, Meta enables researchers who lack access to massive amounts of infrastructure to study and experiment with these advanced systems.
The company makes LLaMA available in several sizes, including 7B, 13B, 33B, and 65B parameter versions, and trains the larger models on 1.4 trillion tokens. Training these smaller foundation models requires far less computing power, which allows researchers to easily test new approaches, validate others' work, and explore new use cases.
This release addresses a significant barrier in the AI field, as full research access to large language models remains heavily restricted due to the massive resources required to run them. By democratizing access to LLaMA, Meta hopes to help the research community better understand how these models work and mitigate known issues like bias, toxicity, and the generation of misinformation.