naïve-labs
← AI standards
Relationship map 05 / AI infrastructure

Data, models& Compute.

Between a dataset and a deployed model lie quality assurance, training, exchange formats, and runtime environments. The map shows the shared standards along that path.

Data quality and inventory form the foundation; exchange formats carry trained models into different runtime and hardware environments.
interactive standarddata, model, or runtimedata or model flow
Data, models, and compute The map follows datasets through training and model exchange formats into runtime environments and embedded hardware. Model exchange formats ISO/IEC 5259-1data quality ONNXstandardised graphs NNEFframework- and hardware-independent OpenVXportable vision and NN graphs SPDX AI Profileinteroperable AI inventory

Light nodes open a concise description and the corresponding primary source. Actors and technical systems provide orientation.

Further developments.

Web Neural Network API (WebNN) W3C Candidate Recommendation W3C ↗

Status
Maintained by