Bulk imports getting stuck at the IA step

What Wildbook are you working in?
I am working in Seals Wildbook.

What is the entire URL out of the browser, exactly where the error occurred?
https://seals.wildme.org/import.jsp?taskId=10f88fca-c8ca-4f77-89ac-ae09b05fb5fe

Can you describe what the issue is you’re experiencing?
For several of my past bulk imports, they have gotten stuck at the IA detection phase. I click on the send to detection button like normal, but after that nothing happens. In the past detection would take about 5 minutes, but I have given this current import 24hrs and it still is not working. Under computer vision status it says: “detection requests sent (0/4 complete). Each detection job in the queue is currently averaging a turnaround time of 0 minutes.” I am not sure what to take away from that or how to fix this issue.

Can you provide steps on how to reproduce what you’re experiencing?
For me each time I upload a new bulk import this issue happens. I go through the normal steps, but once I click “send to detection” my import gets stuck in stasis so to speak.

Hi @CindyE

Thanks for sending your spreadsheet!

  1. Your first media asset file has a .dng extension, which is not a supported file type. Check our Photography Guidelines help doc for a list of supported file types.
  2. Your media assets are very small (such as 460 × 391 pixels). You’ll want to avoid using images less than 480 pixels but not more than 1600 pixels.
  3. There’s a trailing space after “Phoca” in your spreadsheet. Encounter.genus and Encounter.specificEpithet have to match the species name exactly otherwise Wildbook won’t understand which detection algorithm to run. Deleting the space will fix this.

Once you make the updates to your spreadsheet and images, you can delete the import and upload it again with the updated files and it should work more smoothly.

I did all you suggested and it worked perfectly, thank you!

Follow-up question, how critical is it to have our media assets be smaller than 1600 pixels? We have a fair share of images that are larger than 1600pixels but they are mixed in with ones that are already the correct size. It would be possible, although time consuming to compress all the large images.

1 Like

Generally, images that are too small may have a harder time going through detection and ID due to the lack of detail for matching. Images that are too large will take longer to complete detection and ID, especially if they are part of a bulk import.

1 Like

Got it, thanks again!

1 Like