Meta, the innovative technology company founded by Mark Zuckerberg, Tuesday, announced a groundbreaking collaboration with Microsoft to introduce Llama 2, the next generation of their open-source large language model.
Llama 2 will revolutionize the field of artificial intelligence by offering advanced capabilities for both research and commercial applications, and it will be available for free use.
Meta has a long-standing commitment to open-sourcing their infrastructure and AI work, with notable contributions such as PyTorch, a leading machine learning framework, and various models including Segment Anything, ImageBind, and Dino.
By sharing their advancements with the wider community, Meta has fostered progress across the industry, enabling developers to create innovative solutions.
According to Mark Zuckerberg, “Open source drives innovation because it enables many more developers to build with new technology. It also improves safety and security because when software is open, more people can scrutinize it to identify and fix potential issues.
I believe it would unlock more progress if the ecosystem were more open, which is why we’re open-sourcing Llama 2.”
With the release of Llama 2, Meta is offering pre-trained and fine-tuned models with 7B, 13B, and 70B parameters, providing researchers and developers with an extensive range of options to leverage this powerful language model.
Llama 2 has undergone substantial improvements, including enhanced architecture and training on 40% more data than its predecessor, Llama 1.
The fine-tuned models have been meticulously crafted through the collection of over 1 million human annotations, utilizing supervised fine-tuning and reinforcement learning with human feedback (RLHF), resulting in exceptional levels of safety and quality.
To facilitate accessibility, users can download the Llama 2 models directly. Alternatively, through the strategic partnership with Microsoft, Meta has established a preferred arrangement for accessing the models via Azure, coupled with Microsoft’s robust safety and content tools.
Additionally, an optimized version of Llama 2 will be available for local use on Windows, enabling seamless integration with existing workflows.
“We are thrilled about the possibilities Llama 2 brings to the AI community,” said Zuckerberg.
“We are confident that by offering this state-of-the-art open source language model, developers and researchers worldwide will push the boundaries of innovation and unlock new potential in natural language processing.”
Llama 2 represents another significant stride in Meta’s commitment to advancing the field of artificial intelligence through collaboration and open sourcing. This milestone sets a new precedent for accessibility and fosters a culture of shared knowledge to drive technological progress.