LO-SC: Local-only Split Computing for Accurate Deep Learning on Edge Devices

Official implementation of the paper LO-SC: Local-only Split Computing for Accurate Deep Learning on Edge Devices accepted at the 38th International Conference on VLSI Design (VLSID 2025).

Installation

1. Repository setup:

• $ git clone https://github.com/intelligolabs/LO-SC
• $ cd LO-SC

2. Conda environment setup:

• $ conda create -n lo_sc python=3.10
• $ conda activate lo_sc
• $ pip install -r requirements.txt

Run LO-SC

To run the dimonstrative example of LO-SC, use the file demonstrative_example.ipynb. For a detailed, step-by-step explanation of the code and the mathematical aspects of the proposal, refer to optimizer_explanation.ipynb.

Authors

Luigi Capogrosso¹, Enrico Fraccaroli^1,2, Marco Cristani¹, Franco Fummi¹, Samarjit Chakraborty²

¹ Department of Engineering for Innovation Medicine, University of Verona, Italy

² Department of Computer Science, The University of North Carolina at Chapel Hill, USA

¹ name.surname@univr.it, ² enrifrac@cs.unc.edu, samarjit@cs.unc.edu