The new artificial intelligence model announced by OpenAI at an event in recent weeks “GPT-4o” has changed all the balances in this sector. Because the new artificial intelligence model, can perceive the environment It had a structure. That’s the case, OpenAI rivals They also began to conduct similar studies.

In this context; It was revealed that Meta was also working on a new artificial intelligence model. “chameleonThis new model, called “, transforms different types of data into a set of tokens. This data is then made to work in harmony with each other. are integrated. Chameleon will actually be called multimodal artificial intelligence.

Trained on a massive 4.4 trillion dataset

Chameleon is the product of a very extensive work. In this context; Meta, this artificial intelligence model 4.4 trillion It was trained with a huge library of data such as text, audio and images. Artificial intelligence, trained for 5 million hours on fast GPUs, 34 billion parameters used. We should also point out that OpenAI uses about 1 trillion parameters for training.

Meta’s new artificial intelligence model is still in its early stages. In other words: it has not yet reached the end user. there is no new model. However, this situation is expected to change in the near future. Let’s take a look at Meta’s new artificial intelligence model with Google Gemini and OpenAI GPT-4o How will it compete?…