Discussion

Home     Discussion Board      AI Singapore Trusted Media Challenge      Submission error

dinhlinh

Submission error

posted in   AI Singapore Trusted Media Challenge

2021-08-19 01:26

23  comments

dinhlinh

2021-08-20 02:49

Reply

0
<p>Hi, I run this command on local and already produce the result file.docker run —rm -it —gpus all \ -v $fullpath/local_test/test_input:/data/input \ -v $fullpath/local_test/test_output:/data/output \ det_fake_v0 \ -input /data/input/ -output /data/output/submission.csvBtw, I saw the log that the code already run the prediction, maybe the bug is on the writing file result process. Did you observe any stop in the middle during prediction? Did all of the videos are predicted? </p>

Biendata_staff

Yicheng Z.

2021-08-19 08:44

Reply

0
<p>command=’-input /data/input/ -output /data/output/‘+output_file Please note that we have an ‘output’ parameter, and your code must save the output file as whatever that parameter indicates. Here, ‘output_file’ can be a long string of md5 code.</p>

goldbach

2021-08-20 05:01

Reply

0
<p>Hi, I am also hitting the same error since 2 weeks ago. I’ve implemented as per the sample submission repo a main.py file that takes in -output and -input parameters via argparse as suggested by the sample submission repo. This was tested to work locally with <code>docker run -it -v /test_input:/data/input -v /test_output:/data/output —gpus all —rm image_name:latest -input /data/input/ -output /data/output/arbitraryfilename.csv</code>, but did not work on the evaluation server. Additionally, to debug this, I’ve also tried a check in case the -output parameter is in terms of /data/output/some/path/to/directory/anysubmissionname.csv to create the necessary directory and paths, and also implemented the csv saving via the pandas library to align with the sample submission. This was tested to work locally with <code>docker run -it -v /test_input:/data/input -v /test_output:/data/output —gpus all —rm image_name:latest -input /data/input/ -output /data/output/arbitrarydir/arbitraryfilename.csv</code>, but did not work on the evaluation server. Appreciate any clarifications that the organisers can provide. Thank you.</p>

dinhlinh

2021-08-24 10:55

Reply

0
<p>Hi, how about the status now? Could you please clarify soon for me?</p>

Biendata_staff

Yicheng Z.

2021-08-27 02:48

Reply

0
<p>Please refer to our submission guideline <a href="https://github.com/AISG-Technology-Team/Trusted-Media-Challenge-Submission-Guide">https://github.com/AISG-Technology-Team/Trusted-Media-Challenge-Submission-Guide</a> to see how to correctly output a result file.</p>
  • dinhlinh reply Yicheng Z.

    2021-08-28 02:10

    Reply

    0
    <p>Hi, I already tested locally successfully, May I know if there’s an interruption during predicting videos, are there any error log shown? Cause I see no printing log shown, but in the previous submission, I see the pytorch warning log.<br>Also, the message “ the message no result file found” does that mean the prediction was finished and just that there’s no result file found?</p>