Meta 与苹果联手:让 Muse AI 模型在 Mac 上跑得更好Meta and Apple team up to make Muse AI models run better on Macs


AI free token/AI免费积分
Meta 与苹果联手:让 Muse AI 模型在 Mac 上跑得更好

Meta 今天发布了 Muse Spark Open AI 模型,并称其为“迄今为止美国能力最强的开放权重模型”。更值得注意的是,Meta 同时表示这是同类模型中第一个可以“使用两台苹果 M5 Ultra Mac Studio 搭配最新 MLX 框架在桌面端运行”的模型。

Meta 与苹果这对宿敌此次为发布展开合作,并公开互相站台。Meta CEO 马克·扎克伯格在 X 上表示:“Muse Spark Open 权重今日发布——借助与 John Ternus 和苹果的合作,它在新款 Mac Studio 上运行得非常出色。”苹果新任 CEO John Ternus 回复道:“祝贺团队!能以如此高的效率在桌面上直接运行 Muse Spark Open 这种量级和规模的模型,真是太棒了。Mac Studio 不断拓展端侧 AI 的边界。”Meta 首席 AI 官 Alexandr Wang 对 Axios 表示:“很高兴看到爸爸妈妈又和睦相处了”,他同时透露两家公司已合作数月。

据介绍,这款新模型是“原生多模态模型”,总参数量达 8160 亿,在长程任务、复杂软件工程和多步推理方面表现出色。苹果也发布了一篇关于 MLX 的技术文章,介绍这一面向 Apple Silicon 的机器学习数组框架,以及 Muse Spark Open 和另外两个开放模型。

两家公司的联手,意味着 Mac Studio 正被定位为运行超大开放权重模型的桌面级 AI 工作站。对开发者而言,能用本地硬件跑动 8160 亿参数模型,将显著降低对云端 API 的依赖。

本文来源:https://www.theverge.com/news/1012293/meta-apple-muse-spark-open-ai-model-mlx
若有版权问题,请来信告知:13951431913#139.com 或在右方提交评论。

Meta announced the Muse Spark Open AI model today, claiming that it is the "most capable US open-weight model to date." But the company also says that it is "the first model in its class" that can "run on your desktop using two Apple M5 Ultra Mac Studios with the latest MLX framework."

Meta and Apple, typically rivals, worked together on the launch and are publicly supporting each other.

"Muse Spark Open -- weights releasing today -- runs great on new Mac Studios thanks to work with John Ternus and Apple," Meta CEO Mark Zuckerberg said on X.

Meta and Apple team up to make Muse AI models run better on Macs

"Congrats to the team!" new Apple CEO John Ternus said in a reply. "Amazing to run a model of Muse Spark Open's caliber and scale this efficiently, right from your desk. Mac Studio keeps pushing the boundaries of on-device AI!"

"We are excited for mommy and daddy to be playing nice again," Meta chief AI officer Alexandr Wang said to Axios, who also noted that the companies have been collaborating over several months.

The new model is a "natively multimodal model" with 816 billion total parameters that "excels at long-horizon tasks, complex software engineering, and multi-step reasoning," according to Meta. Apple also published a new technical post about MLX, an "array framework" for ML tasks on Apple Silicon, Muse Spark Open, and two other open models.

The tie-up means Mac Studio is being positioned as a desktop-class AI workstation for running very large open-weight models. For developers, being able to run an 816-billion-parameter model on local hardware will significantly reduce dependence on cloud APIs.

暂无评论,快来发表第一条评论吧!No comments yet. Be the first!

📮 需求咨询📮 Inquiry