After OpenAI’s ChatGPT, Google unveiled BARD. Now Meta has also released its new large language model called LLaMA (Large Language Model Meta AI). Meta claims that the LLM is designed to help researchers advance their work.
Meta’s LLaMA is the ‘smaller foundation model’ that requires less computing power and resources. Unlike ChatGPT or BARD, LLaMA is not a chatbot but a research tool to help experts who don’t have access to large amounts of infrastructure to study language models. It will be available in multiple sizes (7B, 13B, 33B, and 65B parameters). LLaMA will be available under non-commercial license to researchers, civil society, and academia.
Also see: AI tools directory
“We believe that the entire AI community — academic researchers, civil society, policymakers, and industry — must work together to develop clear guidelines around responsible AI in general and responsible large language models in particular,” the company wrote in its post.
In a detailed blog post, Meta also included a link to a full research paper and its Github model card. In the research paper, Meta noted that LLaMA-13B outperformed OpenAI’s GPT-3 (175B) on most benchmarks. Once the LLaMA-13B is fully trained, it can be really helpful for small businesses. The company further claims that LLaMA-65B is competitive to DeepMind’s Chinchilla70B and Google’s PaLM-540B.
Since OpenAI released ChatGPT, everyone was talking about AI and machine learning tools. Microsoft also integrated it into Bing and other Microsoft tools to be an early adopter of AI. Google also released Bard. So, Meta also wanted to show the world that it also works in AI. If you look at the history, Meta had multiple failed attempts. It released a chatbot built on LLM – BlenderBot 3, but it couldn’t do well. The company also released Galactica, a model specifically designed for scientific research. Galactica also met the same fate. It will be interesting to see how Meta’s recent attempt turns out.
