How Llama 2 cements Meta's position as AI contender
Meta surprised the world by open-sourcing its AI model Llama. However, it was limited to academicians and researchers. The reception Llama received motivated the company to work on its successor, and it is here now. Dubbed Llama 2, it is a family of AI models intended for both commercial and research purposes. Can it cement Meta's position in the AI race? Let's see.
Why does this story matter?
Meta entered the AI world later than its rivals, Google and OpenAI. However, in the past few months, the company has released multiple AI models for various purposes, including image and music generation. The original Llama signaled Meta's entry into the generative AI race. However, it was a small model compared to others. With Llama 2, Meta hopes to take a leap.
Llama 2 has multiple variants
Unlike Llama, which was only available on request, Llama 2 is free for research and commercial use. It comes in two flavors, Llama 2 and Llama 2-Chat. They are further divided into sub-categories based on the parameters used. There are seven billion, 13 billion, and 70 billion variants. Comparatively, Google's PaLM 2 has 340 billion parameters.
It was trained on 2 trillion tokens
Llama 2 was trained on two trillion tokens and has a context window of 4,096 tokens. It is clear that Meta's new range of models cannot challenge OpenAI's GPT-4 (trillions of tokens) and PaLM 2 (3.6 trillion tokens). However, it can certainly hold its fort for an open-source model. Llama 2's fine-tuned models were trained on over one million human annotations.
Companies will be able to customize Llama 2
Meta has launched Llama 2 in partnership with Microsoft. It is primarily targeting enterprises with Llama 2. Companies will be able to customize the model according to their needs. Meanwhile, outsiders will have a chance to check the models for biases, inaccuracies, and other flaws. It is available through Azure AI, Amazon Web Services, and Hugging Face.
Llama 2's open-source nature will help AI research
Meta's commitment to openness is exciting, as it allows researchers to have an in-depth look at AI ethics and biases. It also helps the company stand out in the ever-increasing AI crowd. Researchers can check Llama 2 for security flaws too, making it potentially the safest AI model. However, Llama 2 could be used for devious purposes as it is free for all.
