Suggestions on Data labelling in DU
I have labelled more than 18 + documents , but I noticed that one of the documents has incorrect field labels. If I correct the labels for that specific document, what would be the best approach to follow after that ??
It’s recommended to export the data before relabeling a document since the pipeline will retrain on all labeled data, including the corrected document. Exporting ensures your changes are reflected in the dataset used for your model or analysis.
After relabeling, you should rerun the pipeline to retrain it on the updated data, ensuring accurate results. While skipping export/import may work for minor changes, it’s best practice to do so for consistent training on the latest data.
This topic was automatically closed 3 days after the last reply. New replies are no longer allowed.