📄 中文摘要
在尝试在 MacBook Air 上本地运行 Qwen 时，用户常常会遇到内存不足的错误，尤其是当设备只有 8 或 16GB 的统一内存时。大型语言模型在消费级硬件上运行似乎不可能，但通过理解量化技术，可以找到解决方案。模型在全精度（FP16）下的内存占用大约为参数数量乘以 2 字节，这意味着 Qwen2.5-7B 模型在 FP16 下的内存需求约为 14GB，超出了大多数 MacBook Air 的内存限制。了解这一点后，可以采取相应的措施来优化模型的运行。
为什么 Qwen 无法在你的 MacBook Air 上运行（以及如何修复它）
出处: Why Qwen Won't Run on Your MacBook Air (and How to Fix It)
发布: 2026年3月29日
📄 English Summary
Why Qwen Won't Run on Your MacBook Air (and How to Fix It)
Users often encounter out-of-memory errors when trying to run Qwen locally on a MacBook Air, especially with only 8 or 16GB of unified memory. Running large language models on consumer hardware seems impossible, but understanding quantization can provide solutions. The memory footprint of a model in full precision (FP16) is approximately the number of parameters multiplied by 2 bytes, which means the Qwen2.5-7B model requires about 14GB of memory in FP16, exceeding the limits of most MacBook Air devices. Once this is understood, steps can be taken to optimize the model's performance.