ML Coverage

ML Life Cycle

Data Labeling/Annotations

Artifacts in ML

• Artifacts ; ML-Ops 에서 생성 혹은 소비되는 Entities
• Artifacts are interdependent.
• Each artifact follows its own life cycle, requiring additional tools and workflows, but needs coordination.
• Records ; Code를 만들기 위한 resources (ex. requirements) 또는 코드의 실행 결과 (ex. test results)

DevOps의 라이프사이클

• DevOps = Development + Test + Operations of software into one integrated workflow

• Use agile principles, best practices, and tools

• DevOps is basis of ML-Ops

• MLOps is a set of best practices that helps manage the creation and use of ML artifacts

• Extend DevOps to manage ML ; Integrate Data Engineering and Model Development into software developing cycle.

MLOps의 라이프사이클