Performance Models
Performance Models support the AI-SPRINT design and runtime components in selecting an appropriate configuration
Performance Models support the AI-SPRINT design and runtime components in selecting an appropriate configuration to: 1) avoid applications performance violations, 2) avoid under or overestimation of continuum resource utilisation, and 3) predict the execution time of Deep Learning components on a target configuration. The best regression model is built for each task, and then used to predict the execution time of inference components/pipelines or training jobs to support the selection of the most appropriate system configuration for executing them, fulfilling QoS requirements while minimising operational costs.
The performance models have been integrated and made available through other AI-SPRINT tools, such as SPACE4AI-D, SPACE4AI-R, and the AI Models Architecture Search (POPNAS).
For any given application, the tool provides as output the relative machine learning models as pickle files and detailed information on the accuracy achieved by the candidate performance models, the set of features selected and the hyper-parameters settin