IEEE P2941.2

This standard defines a set of application programming interfaces (APIs) that can be used on different deep learning inference engines. The interfaces include parameter reading, model compilation optimization, operator registration, thread management, input/output data acquisition, inference instance creation and inference execution.