mobile-ml-benchmark

This repository contains inference latency measurements for 102 real-world CNNs, 1000 synthetic CNNs, 69 real-world ViTs and 1000 synthetic ViTs across 174 diverse experimental environments on mobile platforms, accounting for critical factors affecting inference latency, including hardware heterogeneity, data representations and ML frameworks.

Our paper can be found at:

A Benchmark for ML Inference Latency on Mobile Devices

@inproceedings{li2024benchmark,
  title={A Benchmark for ML Inference Latency on Mobile Devices},
  author={Li, Zhuojin and Paolieri, Marco and Golubchik, Leana},
  booktitle={Proceedings of the 7th International Workshop on Edge Systems, Analytics and Networking},
  pages={31--36},
  year={2024}
}

Table of Contents

• Inference latency measurements for CNNs
• Inference latency measurements for Vision Transformers (ViTs)