Mark Zuckerberg’s Meta Platforms said Friday it was releasing a recent large language model based on artificial intelligence geared toward the research community, becoming the newest company to hitch the AI race.
The battle to dominate the AI technology space, which until recently existed within the background, kicked off late last 12 months with the launch of Microsoft-backed OpenAI’s ChatGPT and prompted tech heavyweights from Alphabet to China’s Baidu to create their very own offerings.
Meta’s LLaMA, short for Large Language Model Meta AI, might be available under non-commercial license to researchers and entities affiliated with government, civil society, and academia, it said in a blog.
The corporate will make available the underlying code for users to tweak the model and use it for research-related use cases. The model, which Meta said requires “far less” computing power, is trained on 20 languages with a deal with those with Latin and Cyrillic alphabets.
“Meta’s announcement today appears to be a step in testing their generative AI capabilities so that they can implement them into their products in the long run,” said Gil Luria, senior software analyst at D.A. Davidson.
“Generative AI is a recent application of AI that Meta has less experience with, but is clearly vital for the long run of their business.”
AI has emerged as a brilliant spot for investments within the tech industry, whose slowing growth has led to widespread layoffs and a cutback on experimental bets. Microsoft, Baidu and Alphabet’s Google, meanwhile, are incorporating their respective advanced AI language engines into more mass products like search.
Meta in May last 12 months released large language model OPT-175B, also geared toward researchers, which formed the premise of a recent iteration of its chatbot BlenderBot.
It later launched a model called Galactica, which it said could write scientific articles and solve math problems, but its demo was later pulled down since it repeatedly generated authoritative-sounding content.