I am able to label the documents and train the ML model based on this. Which is working fine in when tested from Studio.
Now I want to retrain the same model via Business user i.e. using ML Extractor Trainer activity.
Once user validation is done, I can see a new folder created in the desired dataset (the new folder is fine-tune) but now I don’t know how to import these documents in document manager or how to proceed with training the model with newly uploaded documents?
Manually label documents in Document Manager (e.g. 20 documents)
Test it via Studio which should work fine.
New format of documents comes (e.g. 2 documents with a new format)
Business User validates and corrects the extraction result from Robot
Via ML Extractor Trainer activity, label details including documents get uploaded into AI Center> Dataset folder.
From Document Manager> Click on Export> Click on Schedule> then Schedule the export
Scheduling export automatically picks up the labeled documents from fine-tune folder
Schedule picks all the previously labeled documents as well. So the total documents picked up by schedule is 22 i.e. 20 (previously labeled) + 2 (newly labeled via user validation).
Once the export is completed a new folder is created in the Dataset folder.
Now we can create a training pipeline for the ML model and will choose the new dataset folder