Thanks for sending those over! I’ll let you know once those have been reset. Zebra is struggling with a 900 job backlog right now so it may be a little while.
Thank you… please be hurry…I am really anxious to have the result. My whole phd relies on it, and I don’t have much time. Having the result of individual recognition is the first step to do everything else. After the recognition, there will be step 2 and 3 and 10. Yet I have only half a year to finish everything…Please really help me to get the result as soon as possible, I am crying for it.
Sincerely,
Bing
Hi @xie.bing. I wanted to give you a status update so far:
We were able to find a new bug in Zebra Codex based on the list of sighting errors you provided. Those sighting pages now load correctly and matches are currently processing.
3 sightings have broken images in the match results, which we’re working on fixing. They’re rows 31, 41, and 53 in your spreadsheet.:
https://zebra.wildme.org/sightings/7342d5b0-2105-4a2d-8759-8ed3df36c192
https://zebra.wildme.org/sightings/d9c09f3a-8044-4b6b-8639-bdb6b5ff4c82
https://zebra.wildme.org/sightings/b602f063-3676-4f7d-a3b8-4f8375f41821
53 of the sightings currently show “no match candidates”. We’re also working on those because it looks like those are skipping detection since the results come back immediately instead of joining the queue.
Great 
. Thank you so much for trying to solve the problem! Hope to see the result correctly shown soon. After we succeed on the 100 sightings I already sent, there are another 2800 sightings to be dealt with, resetting or so… That would be quite a bunch of work. Thanks in very advance!
Sincerely,
Bing
1 Like
To clarify, we don’t actually have a way to reset them all at once like I previously thought. I had to manually re-run each of the sightings in your list. By doing that, I was able to separate the ones that weren’t working correctly from the ones that properly got queued for identification.
Hi @Anastasia ,
Oh, that’s a bad news… It would be a lot of work from your side to finish 2900 sightings. Shall I send you now the other 2800 sightings? Or shall I just wait to see if the 100 sightings I already sent would work?
Thanks again in advance
Sincerely,
Bing
I can’t manually re-run those 2800 sightings for you. You will need to re-run them. The ones that were re-run last week appeared to go in the detection queue correctly, so go ahead and start doing that.
If you find any that complete detection right away but don’t actually show any match candidates, send those sighting links to us to troubleshoot.
OK, got it.
When I check the 100 sightings I sent back, I can see some of them are good with candidates now, some of them are without any candidates, and some of them are queued behind 0 jobs. Are you still working on them, so I should just wait?
I have tried rerun around 1/5 of the other 2800 sightings, almost all of them are without any candidate. Maybe it’s better for me rerun all other sightings after the 100 sightings are good, which means the problem has been solved from your side. So that I don’t find all the 2800 without any candidate, and need to send back all of them to you.
Sincerely,
Bing
I’m going to email you a spreadsheet that lists the status of each after re-running them. I’ve asked for a status update from the devs on the ones with no match results and broken images. These need to be addressed manually, so it’s taking some time.
That’s odd. What happens if you try to re-run one of these sightings?
Edited to add: I see this message when I tested the sightings with no match candidates. I checked the queue itself and these appear to be processing normally. I’m not sure why Codex isn’t reporting the number of pending jobs, but I’m adding it to our bug report list.
As of right now there are 84 jobs in the queue, but that number will go up slightly as I re-run those 53 without matches. Apparently the system thought they were actively in the ID queue and was ignoring our previous attempts to re-run them. Hopefully these will have match candidates when they finish.
If all of those look good, restart matches on your other sightings (from the 2800 you mentioned). Send us a list of any that still have broken images, no match results, or any other issues after to re-run them.
Hi @Anastasia ,
Thank you for updating.
https://zebra.wildme.org/sightings/84fe9897-08e7-4c99-a77d-100a69b24833
After I re-run the one above, it goes to “queued behind 0 jobs” once again.
OK. Thank you. I will wait and see.
One more question, how many candidates will show up for a sighting? Usually I see 20 candidates when it is good. However, when I check the 100 sightings with some candidates now, most of them are with less number of candidates than 20. Some of them are even with 1 or 2. Moreover, some candidates shown only 0 score, which is weird.
Sincerely,
Bing
There are 161 jobs in the queue right now. I mentioned this above, but I added the missing job queue number to our bug reports.
I’m waiting to hear back from my teammates on this, but I don’t think Codex has a specific number of matches that need to display the same way that Wildbook does.
Great! Thank you! I will wait for the result 
I confirmed that there is no minimum number of matches for Hotspotter. So results like no matches, 1 match, or 2 matches is normal. Like we’ve discussed upthread, no matches only becomes a potential issue if it appears repeatedly across multiple sightings.
Hi @xie.bing
We identified two separate bugs that have been causing in broken images match results in Codex. One affects candidate images and the other affects annotation images. They’re currently being worked on, but I don’t know yet how long it will take to resolve. I just wanted to give you an update.
Hi @Anastasia ,
What does it mean the CDX-54? I can’t check the web through the link, saying I have no access.
Sincerely,
Bing
That’s our internal ticket number for tracking the issue. I should have clarified that it’s only visible to Wild Me employees. The reason I posted it here is so that it’s easier for our team to update Community posts on the status of bug reports if the ticket is linked in the topic.
Got it, thank you. It’s great to know
Thanks for your patience. CDX-54 is now resolved with the latest Codex 2.0 release: Codex 2.0 release notes