dAIEdge-VLab: A milestone in our mission

At dAIEdge, we continue to take giant strides towards becoming a beacon of AI in Europe. Today, we are excited to introduce dAIEdge-VLab, a virtual environment designed to revolutionise the way AI models are evaluated on edge devices.

In this initial phase, access will be restricted to partners, relevant organisations, and researchers, including participants in the open call, as necessary.

  • What is dAIEdge-VLab?

dAIEdge-VLab is a virtual environment designed to compare and assess the performance of AI models across various targets. It provides users with an intuitive interface to configure tests, visualise results, and compare models across different environments and objectives. All of this takes place within a secure setting, where data and models are handled with the highest level of protection.

dAIEdge-VLab enables users to evaluate the performance of their models on different devices or inference engines. Whether the testing environment involves custom or random data, the process is swift and efficient.

Security is a priority. Uploaded data and models are automatically deleted once the evaluation process is complete. Additionally, results remain accessible only during the local session, ensuring confidentiality.

Once the evaluation and analysis are complete, users can review and compare results obtained across different devices and inference engines.

  • How dAIEdge-VLab works?

It features a collaborative architecture comprising several key components:

  1. User Interface

Accessible via a web browser, this interface allows users to upload pre-trained models and select the targets for deployment and evaluation.

  1. Host

A Linux server that coordinates communication with the target devices. It employs GitLab Runners and specially designed Docker images to ensure seamless execution of processes.

  1. Targets

The target devices where models are executed. These can be full operating systems or bare-metal hardware, tailored to the user’s needs.

  1. Target Repositories

These store essential files such as Docker images and scripts required to deploy models on the targets.

 

  • Discover it!

Access the platform and start testing today! A video tutorial is also available to help you explore the platform’s features.

 

Subscribe to our newsletter and be the first to receive updates as we expand access: