Hi I have a project with s3 storage connected for import and export of task data (import is importing blobs). On syncing and inspecting the data I frequently decide that I want to remove a task because that particular image isn’t good quality. However, I have data continuously coming into this bucket, so I have to intermittently resync, how do I prevent deleted tasks being recreated on sync (or being recreated with a different id)? Skipping the task and leaving it as a task that has been marked as skipped seems like a bad proxy, because I want annotators to use the skip column to inform me that there is data I potentially have to review
Hello,
You can add a checkbox to your labeling configuration to indicate that this task contains incorrect data and should be excluded.
Label Studio was designed with the Skip button to ignore tasks, including cases where your data is incorrect. You might consider trying Label Studio Enterprise as it includes review workflows.
Is it possible to preselect a number of annotations and skip them from the data manager? I only see the ability to skip via entering the annotation view. It would be significantly more convenient to mass select and skip if I can. In general I am looking for a way to essentially filter the data prior to handing it off to my annotation team, but the only way I see at the moment treats me like another annotator. I’m hesitant to simply skip because that will mess up the agreement matrices and other tools that will then consider my skip as “An annotator took a look at this image and found nothing.” But really im trying to say, “This image is not worth annotating, please skip it.”