Meta, the tech giant behind Facebook, has announced the release of Llama 2, the next generation of its open-source artificial intelligence (AI) large language model. Llama 2 is available for both research and commercial use and will be distributed through Microsoft's Azure cloud service. This move is part of the expanding partnership between Meta and Microsoft.
Llama 2 is said to be competitive with models such as OpenAI's ChatGPT and Google's Bard chatbot. Developers using Microsoft Azure can utilize cloud-native tools for content filtering and safety features when building with Llama 2. Additionally, Llama 2 is available through Amazon Web Services and Hugging Face, among other providers.
Meta CEO Mark Zuckerberg highlighted the importance of open-source technology in driving innovation and improving safety and security. According to Zuckerberg, Llama 2 was pre-trained on 40% more data than its predecessor and includes architectural improvements. The fine-tuned models of Llama 2 underwent supervised fine-tuning and reinforcement learning with human feedback, resulting in leading results on safety and quality.
Meta believes that an open approach to AI model development promotes transparency and access. The company has subjected its models to red-teaming and has implemented responsible use guidelines and evaluation methods. Meta aims to address the risks associated with AI while providing resources to support responsible usage of Llama 2.
Microsoft's support for Llama 2, despite its investment in OpenAI, reflects the company's strategy of offering developers a choice in the types of models they use to extend its position as a leading cloud platform for AI work.
In summary, Meta's release of Llama 2, an open-source AI language model, in partnership with Microsoft's Azure cloud service, aims to drive innovation and improve safety and security in AI development. The availability of Llama 2 through multiple providers offers developers a choice in utilizing this advanced language model.