MLflowを使う場合の選択肢
1. Managed MLflow by Databricks
a. Notebook integration
b. Scalability
c. Role-based access control
2. Open Source MLflow
a.
自分でホストする必要あり
b. Notebook
も別途構築の必要あり
c.
アクセスコントロールはサポートされていない
本番運用を考慮すると、
Managed
にしたほうが楽+
ML
の中身に集中はできそう
次回以降:
SageMaker Pipelines: the first purpose-built, easy-to-use continuous integration and
continuous delivery (CI/CD) service for machine learning (ML)
Components:
1. pipelines
2. model registry
3. and projects.
https://aws.amazon.com/blogs/machine-learning/building-automating-managing-and-sc
aling-ml-workflows-using-amazon-sagemaker-pipelines/