Meta announces new AI language model , LLaMA

Feb 26

Microsoft, Google and others are battling it out over AI, which could determine the rise and fall of these companies in the next decade and beyond, and how can Meta miss out on the AI market?

Meta announced on Friday that it will launch a new large language model based on artificial intelligence (AI) for the research community, making it an official entry into the field.

LLaMA stands for "Large Language Model Meta AI" and is available under non-commercial licence to government, community and non-profit organisations. The smallest model, LLaMA 7B, has been trained with 1 trillion tokens.

And this training process, in which they chose 20 languages to work with, and some to reduce cases of bias, vicious speech, false speech, etc., is also a challenge Meta is facing.

Among other things, Meta says that LLaMA-13B outperforms the GPT-3 (175B) developed by Open AI in all aspects, while LLaMA-65B is able to compete with Google's Chinchilla70B and PaLM-540B.

As things stand now, LLaMA is still in the process of being made available to the needy, and it will be interesting to see when it will be available to the general public.

Source: Facebook

Another competitor joins in the AI battle, do you think Meta will be able to stand out? Please your thoughts in the comment section below!

All comments (5)
No contentNothing here, please try again later.