Developper Documentation

How to use this documentation

This documentation is divided into three main sections: dAIEdge-VLab, Host, and Targets. Each section contains the necessary information to use the dAIEdge-VLab.

👾

Tip: To develop a new target faster, look at the mandatory guidelines. This will help you to create a target that is compatible with the dAIEdge-VLab with the minimum effort.

ℹ️

Note: The documentation is still under development. If you have any questions or need help, please contact the dAIEdge-VLab team.

Summary

dAIEdge-VLab Host Targets

Introduction

This section explains the interaction between the different components of the dAIEdge-VLab. The dAIEdge-VLab is composed of a Remote server (also referred as dAIEdge-VLab), the Host and the Targets.

The users access the dAIEdge-VLab through a web interface. The user can then upload a pre-trained model and select the target on which the model will be deployed and benchmarked.

The Host is a Linux server that communicates with the target(s). It is based on the use of GitLab Runners. The Host uses the docker image created alongside the target repository.
The target is the device on which the model will be deployed. The target can be any device that can run a model, either on bare metal or on an OS.
The target repository contains the necessary files to deploy the model on the target. The repository contains the docker image and the scripts to deploy the model on the target.
The dAIEdge-VLab is the link between the users and the Hosts. The dAIEdge-VLab is based on the use of GitLab CI/CD pipelines that are triggered by the user’s actions.