About ML Skill training on AI Center

[Long_Nguyen_Hong_CMCTSSG (TonyStark)]

When I finished creating ML Skill on AI Center with trained materials. When running the robot, there are new untrained files but the same format will produce the output that is not what I wanted. So I had Validation Extraction to get the results I wanted. What do I need to do so that the Human Validated Data section is saved to the AI Center for the next rerun the results are more accurate. Thanks

[sharon.palawandram]

If you want to add validated files into the ML model you have created, you can add a train extractor scope which will save the files validated in action center to a dataset for retraining. Seems like you have added those steps, thus it should add the action center files to the dataset for retraining.

Train Extractors Scope

[Long_Nguyen_Hong_CMCTSSG (TonyStark)]

but after using Train Extractors, human validated data is saved to Dataset in Documents and Metadata folder, ML Skill has auto update enabled but when running the same file again, the data is not changed as expected

[Anil_G (Anil Gorthi)]

Hi @Long_Nguyen_Hong_CMCTSSG

Is the pipeline also scheduled to run again?

Without running the pipeline again or without scheduling it ML skill would not get updated

Cheers

[Long_Nguyen_Hong_CMCTSSG (TonyStark)]

Oh! Thank you very much