Meta and Microsoft launch open-source artificial intelligence model "Llama 2"

 Meta and Microsoft Collaborate to Launch Open-Source Language Model "Llama 2"


Microsoft


On July 18th, Meta and Microsoft announced their collaboration to launch the open-source language model, "Llama 2," which will later appear on the Microsoft Windows operating system and the Azure cloud computing platform.

The two teams announced the partnership, stating that Llama 2 is now available for free for research and commercial use, and work is being done to improve it to also be runnable on Windows.





Meta claimed that Llama 2 has been trained on a larger amount – 40% more – of publicly available data sources on the internet, and can process twice the amount of data that Llama 1 was able to process.


The company also stated that Llama 2 outperforms many competing open-source large language models (LLMs) when it comes to encoding, efficiency, and inference, as well as performance in knowledge tests. However, Meta acknowledged that Llama is not as efficient as its competitors, such as OpenAI's GPT-4, according to one of their research papers.

In an Instagram post on July 18th, Meta CEO Mark Zuckerberg stated that Llama 2 "gives researchers and companies access to the next generation of the company's language model and use it for building and development."

Meta claimed to be "surprised" by the high demand for Llama 1 after its limited release in February, receiving over 100,000 requests for access to the model.

However, the model was quickly leaked online by a user on the 4chan website.

Nevertheless, the numbers seen by Llama 1 were far from ChatGPT's numbers, which reached an estimated 100 million users or more during the first three months, according to a Reuters report released in February.

Through this partnership, Microsoft will become a supporter of some of the biggest players in the field of artificial intelligence, as the company has invested $13 million cumulatively in OpenAI throughout 2023, according to a Fortune magazine report in January.

On the other hand, two members of the US Senate criticized Meta's decision to make an open-source Llama model, claiming that the "minimal" protection included in the first version of Llama opened the doors for malicious users to engage in "criminal tasks."


Next Post Previous Post
No Comment
Add Comment
comment url