MLX
Apple 机器学习研究团队为 Apple silicon 写的数组计算框架：统一内存架构让 CPU 与 GPU 共用同一块内存，省掉来回拷贝；API 接近 NumPy，配 mlx-examples 有 LoRA 微调、Whisper、图像生成等可直接跑的样例，是 Mac 本地跑模型的主力栈。 ⭐ GitHub Stars: 29k+ 🔗 仓库: https://github.com/ml-explore/mlx
An array framework from Apple machine learning research built for Apple silicon: the unified memory architecture lets CPU and GPU share one buffer instead of copying, and the API stays close to NumPy. mlx-examples ships runnable LoRA fine-tuning, Whisper and image generation samples, which makes it the default stack for running models locally on a Mac. ⭐ GitHub Stars: 29k+ 🔗 Repository: https://github.com/ml-explore/mlx
📋 Product Info
🎯 Why This Tool
MLX belongs to the Open Source LLMs category on hedirbase. Tagged with: apple-silicon,machine-learning,array-framework,local-llm,open-source.
An array framework from Apple machine learning research built for Apple silicon: the unified memory architecture lets CPU and GPU share one buffer instead of copying, and the API stays close to NumPy.