PrismML brings its tiny LLMs to Qualcomm-powered smart glasses

PrismML 将其微型大语言模型引入高通驱动的智能眼镜

The AI Lab PrismML — founded by Caltech researchers and advised by UC Berkeley’s Ion Stoica — has created a version of its tiny language models for smart glasses running on Qualcomm’s Snapdragon chips. 由加州理工学院研究人员创立、并由加州大学伯克利分校的 Ion Stoica 担任顾问的 AI 实验室 PrismML，已为其微型大语言模型开发了一个适配版本，专门用于搭载高通骁龙芯片的智能眼镜。

On Wednesday, at Qualcomm’s Snapdragon Summit, the chipmaker showcased PrismML’s 1-bit Bonsai LLM, which can be run locally on AI smart glasses built on the Snapdragon AR1 Gen 1 Platform. 周三，在高通骁龙峰会上，这家芯片制造商展示了 PrismML 的 1-bit Bonsai 大语言模型。该模型可以在基于骁龙 AR1 Gen 1 平台的 AI 智能眼镜上本地运行。

As TechCrunch previously reported, PrismML’s claim to fame is that it shrinks larger models substantially (in this case, by 4x), while retaining almost all of their performance on standard benchmarks. 正如 TechCrunch 此前报道的那样，PrismML 的成名之处在于它能大幅压缩大型模型（在此案例中压缩了 4 倍），同时在标准基准测试中几乎保留了其全部性能。

The smart glasses version is a 2-billion-parameter model tuned for vision and language, so wearers can ask what they’re looking at in real time. 这款智能眼镜版本是一个拥有 20 亿参数的模型，针对视觉和语言进行了优化，因此佩戴者可以实时询问他们所看到的内容。

Prism’s larger goal is open-weight AI that runs on devices and makes better use of the computing power they already have. The startup pitches this as an alternative to depending on the privacy promises of proprietary AI labs and their insatiable need for more compute. Prism 的宏大目标是实现可在设备上运行的开放权重 AI，并更好地利用设备现有的计算能力。这家初创公司将其作为一种替代方案，旨在摆脱对专有 AI 实验室隐私承诺的依赖，以及它们对算力无止境的需求。

Releasing a model for Qualcomm’s chip is a step toward that vision. But no smart glasses running PrismML have been announced yet. 为高通芯片发布模型是实现这一愿景的一步。但目前尚未有搭载 PrismML 的智能眼镜产品宣布上市。