Solid-RAG toolbox
Real-world extreme edge prototype to train an LLM (Large Language Model) on personal data in an extreme edge network at people’s home.
Description:
The SOLID-RAG toolbox enables real-world extreme edge deployment of Large Language Models (LLMs) for training on personal data within home environments. It demonstrates a proof of technology for AI-capable edge devices, integrating Solid protocols for consent management and advanced personal data space management. By executing workloads efficiently in an extreme edge network, the toolbox ensures data privacy, minimizes cloud dependency, and optimizes energy consumption, validating the feasibility of training and deploying LLMs directly at the edge.
More information:
Enrichment project: Private and sustainable extreme edge AI applications
Responsible Research Lead: Tanguy Coenen (Strategy lead - imec), Jan Deprez (Innovation specialist - imec) and Matthias Hutsebaut-Buysse (Senior researcher - UAntwperen)
Authors: Maarten Taeymans (imec)
Programming Language: Python
Link to the repository: https://github.com/imec-int/solid-rag-toolbox
Categories: Edge, SOLID
Publications:
Presentation - https://imecinternational.sharepoint.com/:p:/s/PSX/EWvIqDkUTPJIl3yc5lBwYdcB8xzeWxN0vtPf-tv84f5NIA?e=bFAShU