...
- Model management: SANDEEP KUMAR JAISAWAL updated that core functions of model management services are created using golang. Explained the final demo scenario. Can show an initial demo of model management services can be presented in the next meeting.
- Generic Training pipeline: Navaneethan Rajasekaran provided update offline. He was able to create the generic training pipeline and test using kubeflow adapter APIs.
- Support dynamic change of data source: This activity is completed, code is reviewed and merged.
- Training service: Not planned for I release, based on MVP-C slides presented, the training service APIs are yet to be finalized.
- AIMLFW callflows: N.K. Shankaranarayanan gave an overview of the slides prepared. the AIMLFW callflows are not complete. Will have an offline discussion to make an initial set of callflows for AIMLFW. O-RAN alliance has approved AIML workflow as a new function as part of the SMO.
- NTUST: Successfully deployed OSC near-rt ric. Also developing a new pipeline for the AIMLFW for a new usecase. Detailed installation document uploaded for H release. https://wiki.o-ran-sc.org/download/attachments/81297504/OSC%20AI%20ML%20Framework%28Release%20H%29%20Install%20Notes.pdf?api=v2
- Contribution of AIML model monitoring feature: May not be considered as part of I release.
- Energy savings usecase: Navaneethan Rajasekaran is looking into this topic. As first step to look into the analyzing the data.
- Handle restart of VM or system hosting AIMLFW: Yet to start working on this topic.
2023-10-17 (Tuesday)
Recording Link: https://zoom.us/rec/share/nRcyBEOuUcid1TZsfVLLg8PrSzD5Vs3LkCQiUgkeYujvd0pfm-70jHD0DpprKigI.3OStoQDEd4-2fqos
...