Apple has recently unveiled its machine learning framework, MLX, which is specifically designed for Apple silicon chips. The framework is a significant shift in Apple's AI strategy, aiming to democratize machine learning and make it accessible to a broader range of developers and researchers. This move indicates Apple's commitment to making its future operating systems more AI-centric.
MLX offers improved model deployment and training for academics in the Mac, iPad, and iPhone ecosystems. It integrates technologies like LLaMA, LoRA, Whisper, and Stable Diffusion to enhance natural language understanding, image generation, and speech recognition on Apple devices. The framework sets itself apart with features like shared memory storage, composable function transformations, lazy computation, and dynamic calculation graphs.
The framework also boasts features such as familiar APIs, effortless efficiency, lazy computation, dynamic computation graphs, multi-device support, unified memory advantage, and researcher-friendly design. It draws inspiration from established frameworks like NumPy, PyTorch, Jax, and ArrayFire. MLX is available on PyPi and can be installed with the command 'pip install mlx'.
Apple's AI efforts are spearheaded by John Giannandrea and Craig Federighi, with Eddy Cue also involved. The company plans to purchase AI servers in the coming years, but its purchases will be lower than its competitors due to a shortage of Nvidia AI chips. The company is also investing in AI technology, including building its own large language model and revamping Siri.