MLX: The New Open-Source Framework from Apple that Boosts AI Development on M-Series CPUs

Apple is well-known for its innovative approach and complete control over its technology stack. It took a risk a few years ago when it switched to silicon computer chips, the M-series CPUs, for its devices. It recently made another significant breakthrough by introducing MLX, an open-source framework specifically designed to perform machine learning on Apple’s M-series CPUs.

Machine learning is a hot trend in the technology industry, and many AI software development projects use open-source Linux or Microsoft systems. Apple intends to keep up with this trend by providing the most up-to-date tools and capabilities to its thriving developer ecosystem.

What is MLX and Why is it important for Apple?

“MLX” is more than just a technical solution; it is an innovative and user-friendly framework inspired by popular frameworks like PyTorch, Jax, and ArrayFire. It facilitates the training and deployment of AI models on Apple devices without sacrificing performance or compatibility.

It has a unique memory model that allows arrays to exist in shared memory, allowing operations to be performed across multiple device types without duplicating data. This means that developers can use their Mac’s RAM for all tasks without purchasing a separate GPU with a large amount of VRAM. This feature gives developers more flexibility and convenience, particularly those who work on multiple AI projects.

Because of Apple Silicon’s closed ecosystem and lack of compatibility with many open-source development projects and their widely used infrastructure, developing AI software has been difficult.

“Apple has always been a bit of a black box when it comes to AI development. It’s hard to port custom models to their devices, and you have to rely on CoreML, which is not ideal for innovation,” a developer said on Hacker News in a discussion of the announcement.

How MLX compares to other frameworks like PyTorch and CoreML


CoreML is a framework that allows developers to convert and optimize existing machine learning models for Apple devices. However, it is not ideal for developing and deploying machine learning models from scratch, as intended. MLX provides tools for innovation and development within the Apple ecosystem, allowing developers to take advantage of Apple’s hardware’s power and efficiency.

MLX outperformed PyTorch in benchmark tests, outperforming in terms of image generation speeds, especially at larger batch sizes. It is also compatible with tools like Stable Diffusion and OpenAI’s Whisper, representing a significant advancement in AI research and development.

For example, Apple reports that it takes “about 90 seconds to fully generate 16 images with MLX and 50 diffusion steps with classifier free guidance and about 120 for PyTorch”.

With the development of MLX, Apple has reached an important milestone in the advancement of artificial intelligence (AI). This technology not only addresses technical issues but also opens up new avenues for AI and machine learning research and development on Apple devices. The goal of MLX is to make Apple’s platform more appealing and accessible to AI researchers and developers. Furthermore, it promises to provide an exciting and rewarding experience for Apple fans who are interested in AI.

Can you use MLX for Natural Language Processing Tasks?

MLX can be used for natural language processing tasks. It is an easy-to-use and efficient framework for developing and running machine-learning models on Apple’s M-series CPUs. It is capable of performing a variety of machine-learning tasks, including classification, regression, generation, and reinforcement learning.

Natural language processing (NLP) is the science of creating machines that can manipulate human language — or data that looks like human language — in the way it is written, spoken, and organized. Text classification, summarization, text generation, translation, and other tasks are possible with NLP.