Apple just made something new called MLX and it is exclusively for developers.

Apple MLX is accessible through popular open-source libraries like PyPI and GitHub, marking a significant move towards making machine learning more accessible. The framework is designed with users in mind. It is equipped with user-friendly features while maintaining efficiency in training and deploying models.

Apple said that MLX is designed to be easy for users while also being effective for training and using models. It added that MLX is crafted by experts in machine learning, aiming to make it simple for researchers to build on and enhance the framework.

Apple drew inspiration from frameworks like PyTorch, Jax and ArrayFile while developing MLX. The framework is like magic. It uses a special memory trick. It is like making things easier and faster.

The key features of Apple MLX include familiar APIs, dynamic graph construction and multi-device capabilities. The Python API closely follows NumPy. MLX provides a fully featured C++ API mirroring the Python API. Dynamic graph construction has been simplified to offer developers optimized capabilities for debugging.

This move comes amidst the launch of Google’s Gemini AI model, indicating Apple’s quiet innovation in the AI space. While reports surfaced in 2023 about Apple developing its large language model (‘Apple GPT’), the company has remained tight-lipped about such projects.

Apple CEO Tim Cook acknowledged the importance of AI and ML in the company’s products, emphasizing their fundamental role. Speculation about its AI approach has been fueled by increased R&D spending and the company’s emphasis on locally-run AI models leveraging its in-house silicon capabilities. The introduction of MLX suggests Apple’s commitment to advancing AI tools and technologies.

Show
comments