Meta and Microsoft Partner Launch Open-Source LLaMa 2 : Double Context Length, Improved Performance

Meta announces a ‘new and better’ LLaMa 2 with Microsoft

Microsoft has also a collaboration OpenAI which mainly develops LLMs that are closed-source.

Microsoft and Meta have teamed up to create an open-source large-language model similar to ChatGPT. LLaMa 2 was trained using 40% more data than LLaMa 1 which Meta launched in February.

It’s open-source nature sets it apart from competitors such as OpenAI’s ChatGPT or Google’s Bard.

Microsoft is collaborating with Meta which has been an advocate of open-source…

Instagram.

Meta claimed in a press statement that LLaMa 2, which has twice the context length of LLaMa 1, \”outperforms other LLMs such as Falcon and MPT\” regarding reasoning, coding and proficiency tests.

Source:
https://interestingengineering.com/science/meta-new-and-improved-llama-2-partnership-with-microsoft

发表回复

您的邮箱地址不会被公开。 必填项已用 * 标注