Frequently asked questions¶
How long do I need to wait for detection? Identification?¶
Machine learning jobs in Wildbook run in a shared, serial queue. Detection jobs can take between 10 seconds and hours, depending on how much shared ML work is in the queue. Identification jobs generally run only a few minutes but also share resources in the queue and can take longer while waiting for work farther ahead in the queue to complete.
How can I tell if a detection job has completed?¶
You can tell if a detection job has completed by visiting the encounter page. If the images in the gallery have an annotation (dashed green line) around the animal, detection has completed. You will need to refresh the page to see the annotation appear.
How can I tell if an identification job has completed?¶
You can tell if an identification job has completed by clicking the hamburger menu on an image in the encounter gallery and checking for “View Match Results”. If, instead, the menu says “Waiting on results”, the job is still processing. You will need to refresh the page to see the menu option change. If the menu says “Cannot start match” let us know in the Community Forum.
What’s wrong with my match results?¶
“Attempting to fetch results” or “Gave up trying to match” usually means that matching is still in progress. “No match confirmed” or “No match results” means that there are no potential matches for that animal based on the location ID you selected. If you see “Unknown error”, try re-running the match. If that doesn’t produce matches, let us know in the Community Forum.
What should I put in media asset columns in a bulk import?¶
Media asset columns should be named
"Encounter.MediaAsset.X", where X is an actual number, starting with 0
(Encounter.MediaAsset.0). The content of each cell should be the asset filename, including the path to the asset. For example, the
"Encounter.MediaAsset.1" column could contain cells
"january/2.jpg". Check the Photography guidelines for supported image formats.
Why is my bulk import taking so long?¶
Although bulk imports can support up to 1000 encounters in a single spreadsheet, that also makes it slow to process and slow to load. Try to limit each spreadsheet to a few hundred encounters or less.
How do I delete an encounter?¶
Click the Edit button on the Metadata window of an Encounter. Then click Delete Encounter?