Meta is creating a brand new synthetic intelligence mannequin referred to as “Chameleon” to compete with OpenAI’s GPT-4o mannequin. Nonetheless, Chameleon continues to be in its early levels.
OpenAI just lately introduced their new synthetic intelligence mannequin, GPT-4o, at an occasion held up to now few weeks, which has considerably impacted the sector. The brand new AI mannequin can understand its setting, prompting OpenAI’s rivals to begin comparable initiatives.
On this context, it has been revealed that Meta can be engaged on a brand new AI mannequin named “Chameleon.” This mannequin converts varied varieties of information right into a set of tokens, that are then built-in to work harmoniously. Chameleon will probably be often known as a multimodal synthetic intelligence.
Skilled with an enormous dataset of 4.4 trillion
![Meta's New AI Model Chameleon to Challenge GPT-4o by OpenAI](https://metaverseplanet.net/blog/wp-content/uploads/2024/05/Metas-New-AI-Model-Chameleon-to-Challenge-GPT-4o-by-OpenAI.jpeg)
![Meta's New AI Model Chameleon to Challenge GPT-4o by OpenAI](https://metaverseplanet.net/blog/wp-content/uploads/2024/05/Metas-New-AI-Model-Chameleon-to-Challenge-GPT-4o-by-OpenAI.jpeg)
Chameleon is the results of a complete examine. Meta educated this synthetic intelligence mannequin with an intensive library containing 4.4 trillion items of textual content, audio, and pictures. The AI was educated on high-speed GPUs for five million hours, using 34 billion parameters within the course of. Compared, OpenAI makes use of round 1 trillion parameters for its coaching.
Meta’s new AI mannequin continues to be in its early levels, which means it has not but reached the tip consumer. Nonetheless, that is anticipated to vary within the close to future. It will likely be fascinating to see how Meta’s new AI mannequin will compete with Google’s Gemini and OpenAI’s GPT-4o.
You might also like this content material
Observe us on TWITTER (X) and be immediately knowledgeable concerning the newest developments…
Copy URL