Reproduce “Large Language Model Deployment on Resource-Constrained Edge Devices: A Practitioners Survey” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Rahul Maliakkal, Yashasvi Makin, Plawan Rath, Rohit Jain, Ankur Sadhoo
The authors survey deployment of large language models on constrained edge devices by coordinating adaptation and compression, inference methods, runtimes, and hardware. They review quantization, pruning, distillation, parameter-efficient fine-tuning, key–value caching, and speculative decoding, then offer device-class guidance. Their estimate puts a 7B model’s memory footprint at about 4 GB with 4-bit quantization, but no experiment validates deployment on a phone.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.