Same submission but failed in phase 2

Hi, I just replace the query_descriptors.npz in the submission fold which can meet the phase1 requirement, then I submit it in phase2, it returns

+ [[ -f query_descriptors.npz ]]
+ [[ -f reference_descriptors.npz ]]
+ echo 'Validating submission...'
Validating submission...
+ conda run --no-capture-output -n condaenv python /opt/validation.py --query_features query_descriptors.npz --ref_features reference_descriptors.npz --query_metadata /data/query_metadata.csv --ref_metadata /data/reference_metadata.csv
Traceback (most recent call last):
File "/opt/validation.py", line 128, in <module>
main(args)
File "/opt/validation.py", line 117, in main
validate_sorted_ids("query", query_features['video_ids'])
File "/opt/validation.py", line 91, in validate_sorted_ids
raise DataValidationError(
__main__.DataValidationError: Video ids not sorted at index [42033].
ERROR conda.cli.main_run:execute(47): `conda run python /opt/validation.py --query_features query_descriptors.npz --ref_features reference_descriptors.npz --query_metadata /data/query_metadata.csv --ref_metadata /data/reference_metadata.csv` failed. (See above for error)

I checked that I have sorted the ids in the main.py, could you please take a look?

Hey Zihao-

Can you double check that your submitted query_descriptors.npz file contains video IDs in sorted order? Your inference code appears to be generating sorted descriptors in the runtime but your submitted descriptors appear to be out of order.

Thanks,
Chris

Hi Chris:

Thanks for your advice. I found disorder in my query_descriptors.npz. Now I’ve successfully submitted it. Before, I thought the mistake was in main.py

Really appreciate it!

Best regards

Zihao

1 Like