Figure 1
Typical ML lifecycle based upon ISO-23053 (ISO-23053, 2022).
Figure 2
Enhanced Overview of Machine Learning category in ProMetaS (Sherpa et al., 2023).
Figure 3
Sketch of Dataverse-MLflow integration of models.
Figure 6
Flow regime classification using image data.
Figure 7
Workflow Demonstration of Dataverse-MLflow. (1) Image data is collected and uploaded to Dataverse along with metadata. (2) The dataset is cleaned and preprocessed. (3) A suitable model is selected from a collection and uploaded to Dataverse. (4) The shared model is deployed. (5) The performance of the ML model is monitored using predefined metrics.
Figure 4
Incorporation of Metadata from MLflow (left) to Dataverse with ProMetaS schema(right) using DMIL model module.
Figure 5
Sketch of Dataverse-MLflow integration of projects.