Split-Et-Impera is a novel and practical framework that i) determines the set of the best-split points of a neural network based on deep network interpretability principles without performing a tedious try-and-test approach, ii) performs a communication-aware simulation for the rapid evaluation of different neural network rearrangements, and iii) suggests the best match between the quality of service requirements of the application and the performance in terms of accuracy and latency time.
Split-Et-Impera has the following dependencies:
- Sphinx (optional), for documentation generation;
- SCNSL (mandatory), for simulations of networked systems.
SCNSL is a SystemC Network Simulation Library, licensed under the LGPL license. It extends SystemC with capabilities for performing simulations of networked systems. Here is the official repository, and here is the documentation of the library. For further information about SCNSL, check the paper.
- Clone this repo, and we'll call the directory that you cloned as ${SEI_ROOT}.
- Install dependencies. We use Python >= 3.8 and PyTorch >= 1.7.0.
- To use the code from sources in a clean way, it is preferable that you set up a conda enviroment, say:
$ conda create -n sei python=3.8 $ conda activate sei $ conda install pytorch==1.7.0 torchvision==0.8.0 cudatoolkit=11.3 -c pytorch $ pip install -r requirements.txt
The software files in this repository are provided under the MIT License.
Luigi Capogrosso, Federico Cunico, Michele Lora, Marco Cristani, Franco Fummi, Davide Quaglia
Department of Computer Science, University of Verona, Italy
name.surname@univr.it
If you use Split-Et-Impera, please, cite the following paper:
@inproceedings{capogrosso2023split,
title={Split-Et-Impera: A Framework for the Design of Distributed Deep Learning Applications},
author={Capogrosso, Luigi and Cunico, Federico and Lora, Michele and Cristani, Marco and Fummi, Franco and Quaglia, Davide},
booktitle={2023 26th International Symposium on Design and Diagnostics of Electronic Circuits and Systems (DDECS)},
pages={39--44},
year={2023},
organization={IEEE}
}