Description
This facility provides the hardware for the AgrospAI Data Space, which is federated with the agrifoodTEF Data Space and offers, among other things, compute-to-data capabilities and multiple data room configurations for running data processing services, including AI and particularly Deep Learning. The data space is backed by a cluster that, combined, provides 280 CPU cores, 1.7 TB of CPU memory, 6 GPUs providing a total of 288 GB of GPU memory, and 28 TB of high-speed storage.
What could be offered through this facility?
The AgrospAI Data Space Cluster provides a state-of-the-art environment for end-to-end validation of AI and robotic systems, enabling:
(1) Federated compute-to-data operations across AgrospAI and agrifoodTEF Data Spaces, allowing secure access to distributed datasets without compromising data privacy.
(2) Large-scale simulation and real-world benchmarking for AI models and robotic platforms, from autonomous ground vehicles to intelligent manipulators, including performance and behaviour under variable environmental conditions.
(3) Rapid prototyping with real-time feedback loops, supporting iterative development and tuning of algorithms for perception, planning, and control.
(4) Custom algorithm benchmarking on massive agro-industrial datasets, including computer vision, LiDAR, and sensor fusion data, to evaluate accuracy, robustness, and scalability.
(5) Edge-to-cloud performance testing, enabling seamless evaluation of AI workloads across heterogeneous deployments, from embedded robotics platforms to cloud-based AI services.
(6) High-throughput data analytics for deep system insights, identifying bottlenecks, optimising resource usage, and enhancing model interpretability.
(7) Safety, reliability, and compliance testing, ensuring that AI-driven robotic solutions meet the highest standards for operational safety, regulatory compliance, and industrial robustness.
(1) Federated compute-to-data operations across AgrospAI and agrifoodTEF Data Spaces, allowing secure access to distributed datasets without compromising data privacy.
(2) Large-scale simulation and real-world benchmarking for AI models and robotic platforms, from autonomous ground vehicles to intelligent manipulators, including performance and behaviour under variable environmental conditions.
(3) Rapid prototyping with real-time feedback loops, supporting iterative development and tuning of algorithms for perception, planning, and control.
(4) Custom algorithm benchmarking on massive agro-industrial datasets, including computer vision, LiDAR, and sensor fusion data, to evaluate accuracy, robustness, and scalability.
(5) Edge-to-cloud performance testing, enabling seamless evaluation of AI workloads across heterogeneous deployments, from embedded robotics platforms to cloud-based AI services.
(6) High-throughput data analytics for deep system insights, identifying bottlenecks, optimising resource usage, and enhancing model interpretability.
(7) Safety, reliability, and compliance testing, ensuring that AI-driven robotic solutions meet the highest standards for operational safety, regulatory compliance, and industrial robustness.
Type of facility
Fixed
Reconfigurable
Sector
Aquaculture
Arable farming
Forestry
Food processing
Greenhouse
Horticulture
Livestock farming
Tree Crops
Viticulture