Apple's machine learning research team quietly released a framework called MLX for building foundational models.

Many people believe that Apple is cautious in the field of artificial intelligence, but it has quietly released frameworks and model libraries that can run on its chips, and may bring generative AI applications to MacBook.

The company's machine learning research team has released MLX, a machine learning framework where developers can build models that run efficiently on Apple Silicon and the MLX Data deep learning model library. Both can be accessed through open repositories such as GitHub and PyPI.

According to Apple on GitHub, frameworks like PyTorch, Jax, and ArrayFire inspired the design of MLX, with the notable difference of having shared memory, meaning any task running on MLX can work on supported devices (currently CPU and GPU) without moving data. Computerworld reported that MLX aims to provide developers with both ease of use and enough capability to train AI models like Meta's Llama and Stable Diffusion. The framework and model library contribute to driving many AI applications currently on the market.

Apple's machine learning researcher Awni Hannun tweeted that MLX Data is a "framework-agnostic, efficient, and flexible data loading package" that can work with MLX, PyTorch, or Jax frameworks.

Apple has been using artificial intelligence for years, embedding the technology into its products.

However, Apple's main focus is on machine learning and has not ventured into popular generative AI applications pursued by competitors like Microsoft and Google. Apple even avoids using the term "artificial intelligence" in many key presentations.

It is reported that Apple began researching foundational models in September to select models that can be applied in its services.