john-rocky

CoreML-LLM

john-rocky

Run LLMs on Apple devices with CoreML, optimized for Apple Neural Engine + GPU

AI 简介

CoreML-LLM 是一个面向苹果设备的端侧大语言模型(LLM)运行框架,专为 Apple Neural Engine(ANE)优化,兼顾 GPU 协同调度。支持 Gemma 4、Qwen3.5、Qwen3-VL、FunctionGemma、EmbeddingGemma 及 LFM2.5 等多类模型,提供文本生成、多模态理解(图像/视频/音频)、函数调用与嵌入向量生成能力,并具备流式响应、多轮对话和本地缓存复用等特性。适用于对隐私敏感、需离线运行、低功耗或需 GPU 资源留给其他任务(如渲染、AR)的 iOS/macOS 应用场景。

Python
MIT License
170
Stars
19
Forks
3
Watchers
2
Issues

Star 增长

今日0
近 7 天0
近 30 天+7
综合评分44.6
默认分支main