Generative AI (GAI) is all the rage these days, thanks to new chatbots and AI services like OpenAI’s ChatGPT, Microsoft’s Bing Chat, Google’s Bard, and more. And with the passage of time, these technologies are only improving with the addition of new capabilities to cater to more audiences. For those unaware, modern AI chatbots are powered by large language models (LLMs), and today, Meta has announced the availability of its latest LLM, Llama 2.
Meta has tapped Microsoft as its preferred partner for the rollout of Llama 2, its latest free, open-source LLM. As the name suggests, Meta’s Llama 2 is an enhancement of the original release from a few months ago. Meta says that it received lots of interest from researchers regarding the personal and commercial use of Llama. As users began to leverage Llama across a variety of use-cases and provide feedback, Meta was able to fine-tune the model and fix several issues for its latest Llama 2 release.
Following the announcement of a partnership between Meta and Microsoft, several details regarding availability were shared too. Microsoft Azure customers can utilize Llama 2 through the Azure AI model catalog and the LLM has been optimized to run smoothly in local Windows environments too. Additionally, Llama 2 will soon be made available through other cloud platforms and services like AWS and Hugging Face.
With the democratization of AI models becoming more common, Meta has emphasized the importance of safety. To that end, it says that it tested the model in adversarial environments, developed a transparency schematic to highlight known issues and challenges, published a responsible use guide, and authored policies to bar the use of the LLM in certain problematic environments and use-cases. Finally, Meta is also heading an Open Innovation AI Research Community where academics and researchers can contribute to the understanding of responsible development when it comes to LLMs. If all of this interests you, you can download Llama 2 from here.