5 things you should know about Meta’s ChatGPT-rival, AI platform LLaMA
Meta, the parent company of Facebook, announced that its ChatGPT-rival large language model LLaMA will be available for free. Check out 5 things about the newest AI platform in town.
Facebook's parent company Meta might have been late to the AI party, but it has still managed to make a grand entry. Yesterday night, the company announced that its large language platform LLaMA will be made open-source and entirely free for research and commercial purposes, taking a similar approach to OpenAI's ChatGPT. The move was announced during Microsoft's Inspire event, where Meta also revealed its partnership with Microsoft, and extended its support for products such as Azure and Windows. And these are the 5 important things you should know about Meta's AI platform.
1. What is LLaMA?
In February, Meta announced its LLaMA (Large Language Model Meta AI) platform, a large language model built on 65 billion parameters. At the time of its announcement, Meta revealed that the platform was designed to help researchers advance their work in this subfield of generative AI. Acknowledging it as a smaller foundation model, the company claimed that for many researchers trying to understand generative AI, smaller models are more desirable as “it requires far less computing power and resources to test new approaches, validate others' work, and explore new use cases”.
Recently, the company released its LLaMA 2 platform.
2. Why are Meta and Microsoft partnering for LLaMA
At the Microsoft Inspire event, Meta announced its partnership with the Windows maker. Under this collaboration, Meta will extend the capabilities of its LLaMA 2 platform to Microsoft's Azure and Windows. In an announcement, Microsoft said, “Llama 2 is designed to enable developers and organizations to build generative AI-powered tools and experiences…We offer developers choice in the types of models they build on, supporting open and frontier models, and are thrilled to be Meta's preferred partner as they release their new version of Llama 2 to commercial customers for the first time”.
3. What exactly will this partnership mean for developers building on Azure and Windows
Microsoft explained that LLaMA 2 will be part of Azure's AI model catalog. The model is in public preview and will allow developers to eliminate the need to manage all infrastructure dependencies when operationalizing Llama 2. It provides turnkey support for model fine-tuning and evaluation, including powerful optimization techniques such as DeepSpeed and ONNX Runtime, that can significantly enhance the speed of model fine-tuning, mentioned the blog.
4. Why is LLaMA now open-source and free to access
Meta revealed that the LLaMA platform will be open-source and free for research and commercial purposes. Meta claims that the reason behind it is mutual benefits. While the users will be able to access the AI tool and use it, it will also make the LLM more robust and help it advance.
Highlighting the latter, Meta stated, “Opening access to today's AI models means a generation of developers and researchers can stress test them, identifying and solving problems fast, as a community. By seeing how these tools are used by others, our own teams can learn from them, improve those tools, and fix vulnerabilities”.
5. Focus of LLaMA on transparency
Meta is performing a number of steps to ensure LLaMA remains a responsible platform with a focus on transparency and access, the company stated. Some of the exercises the company is undertaking include Red-Teaming Exercises, sharing its model's transparency schematics in its research paper, sharing a user guide, and a fair use policy to restrict any harmful behavior that may come from the use of the AI.