Skip to main content
. 2021 Oct 14;110(11-12):2993–3013. doi: 10.1007/s10994-021-06052-0

Table 1.

Comparison between different frameworks and platforms for managing the ML Lifecycle

Frameworks Usage Init DM TR TS DE UM
Datatron (Datatron 2021) Commercial
Peltarion (Peltarion 2021) Commercial
SELDON (Seldon 2021) Commercial
Algorithmia (Algorithmia 2021) Commercial
5 Analytics (5Analytics 2021) Commercial
craft ai (craft ai 2021) Commercial
KNIME (KNIME 2021) Commercial
valohai (valohai 2021) Commercial
AML (Microsoft 2021) Commercial
SageMaker (SageMaker 2021) Commercial
Tfx Baylor et al. (2017) Internal
FBLearner (FBLearner 2021) Internal
Michelangelo (Michelangelo 2021) Internal
kubeflow (kubeflow 2021) Community
Airflow (airflow 2021) Community × ×
Flyte (Flyte 2021) Community × × ×
NiFi (NiFi 2021) Community × × ×
MLflow Zaharia et al. (2018) Community × ×
Cortex (Cortex 2021) Community × × × ×
JupyterHub (JupyterHub 2021) Community × × ×
ModelHub Miao et al. (2017) Community × × ×
MLife Community

Init System setup and initialization, DM Data Management, TR Model Traning, TS Model Testing, DE Model Deployment, UM Using and Monitoring

(need additional works[]; support[]; nonsupport[×]; unclear[−])