Hello @Sumit Bera , Thanks for your time and patience throughout this issue.
The product team confirmed that this is the current limitation of the service. Unfortunately, this feature does not let you retrain the existing model with new data. The best approach would be to delete the previous model and use the same name for the new model because you will not be able to retrain. However, we understand that this is not an ideal approach, especially if you have already paid to train the previous model.
We recommend training another model with this data and making your production code configurable and version the model in its name so it's easily replaceable. While this may require additional effort.
We apologize for any inconvenience this may cause, and we appreciate your understanding.
I hope this helps.
Regards,
Vasavi
Please remember to "Accept Answer" if any answer/reply helped, so that others in the community facing similar issues can easily find the solution.