Offline training occurs during development before an AI system is placed into production. The model is trained and tested in advance, rather than learning continuously from live production data.
βThe system is trained during the development process before the system is put into production. This is similar in nature to standard software development, where the system is built and tested fully before it is put into production.β ISO/IEC 22989:2023.