Details
-
New Feature
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
None
-
None
Description
We plan to integrate qlib into the submarine after discussion in a weekly online meeting.
The goal of this Jira is
- Run qlib workflow in the notebook
- Run qlib workflow in the experiment (Both single node and multi-node training)
- Support MLflow UI on the workbench (Launch a pod in K8s like tensorboard) since qlib will write underly metadata (metrics and parameters) to mlflow server
- Support different versions of the qlib library in the environment for users to choose from.
Attachments
1.
|
Run qlib workflow in the notebook | Resolved | Kevin Su | |
2.
|
Support MLflow server on Submarine | Resolved | Wen-Chien,Juan | |
3.
|
Frontend support for MLflow UI | Resolved | Wen-Chien,Juan | |
4.
|
Support MLflow model registry | Resolved | Wen-Chien,Juan |