Mobile AI Factory:为移动设备构建专用语言模型的智能系统
Mobile AI Factory: Building Specialized Language Models for Your Mobile Device
Mobile AI Factory 通过 LLM 指挥、设备实测和自适应搜索,自动为特定任务和手机生成微调量化的小语言模型(SLM),在显著减少 GPU 时间的情况下实现或超越大型模型的性能。
Published on Nov 10
Authors:
,
,
Abstract
Mobile AI Factory is an agentic system that automatically builds task- and phone-specific fine-tuned quantized small language models, using an orchestrator LLM, real device measurements, and adaptive search to match or outperform larger models with much less GPU time.
Small language models (SLMs) running on mobile devices can keep user data private and work without connectivity. But they perform well only when fine-tuned for the target task and sized for the target device, a process that today requires repeated expert tuning. We present {\it Mobile AI Factory}, an agentic system that builds a fine-tuned, quantized SLM for a given task and phone. Given a task specification and a target phone, an orchestrator LLM chooses whether to tune hyperparameters, mine real data, or generate synthetic examples, while fixed rules roll back changes that hurt accuracy. The factory searches from small to large models across base models and quantizations, adapts its accuracy goal as it searches, and uses real measurements of the target phone across inference engines and backends. We evaluate Mobile AI Factory on ten tasks and four Android phones. On an unseen testset, its on-device models outperform two five-shot frontier models on five tasks, by up to 0.34, and match them on three more. Its search comes within 0.03 of exhaustively training all 27 model variants on an entity extraction task, while using 19times less GPU time. Phone measurements also matter: the best configuration changes across phones. Without phone measurements, Claude Opus 5.5, given the full hardware and model specifications, misses the app's requirements in 25\% of easy and moderate cases and in 60\% of hard ones. Code is available at: https://github.com/evanly-gh/SLM_Factory.
View arXiv page View PDF GitHub 2 Add to collection
Models citing this paper 4
Datasets citing this paper 0
No dataset linking this paper
Cite arxiv.org/abs/2611.13509 in a dataset README.md to link it from this page.
Spaces citing this paper 0
No Space linking this paper
Cite arxiv.org/abs/2611.13509 in a Space README.md to link it from this page.
Collections including this paper 2
来源: Hugging Face Daily Papers · huggingface.co