Watermark task submission failure: HTTP500

PAN
1

The dry run works fine, but whenever I tried to submit I got errors

TIRA Code Submission:
✓ You are registered and will submit as team xxx to task text-watermarking-panclef-2026.
✓ The dataset spot-check-dataset-20260311-training is available locally.
✓ The code is in a git repository /Users/yy/Desktop/clef26/watermark/clef26-dev.
✓ The git repository /Users/yy/Desktop/clef26/watermark/clef26-dev is clean.
✓ The code is embedded into the docker image submissions-v11-63b30.
✓ The docker image produced valid outputs on the dataset spot-check-dataset-20260311-training. (You can verify them at /var/folders/_s/9wvwt8_n2_n1ygfh_73f84gh0000gn/T/tira-q7yq3zxk/output)
✓ The meta data is uploaded to TIRA.
✓ The Docker image is pushed to TIRA.
Configure code submission in TIRA...
Upload of software failed with error 
<!doctype html>
<html lang="en">
<head>
  <title>Server Error (500)</title>
</head>
<body>
  <h1>Server Error (500)</h1><p></p>
</body>
</html>
 and response code 500.
Traceback (most recent call last):
  File "/Users/yy/miniconda3/lib/python3.10/site-packages/tira/rest_api_client.py", line 317, in add_docker_software
    ret = json.loads(ret)
  File "/Users/yy/miniconda3/lib/python3.10/json/__init__.py", line 346, in loads
    return _default_decoder.decode(s)
  File "/Users/yy/miniconda3/lib/python3.10/json/decoder.py", line 337, in decode
    obj, end = self.raw_decode(s, idx=_w(s, 0).end())
  File "/Users/yy/miniconda3/lib/python3.10/json/decoder.py", line 355, in raw_decode
    raise JSONDecodeError("Expecting value", s, err.value) from None
json.decoder.JSONDecodeError: Expecting value: line 2 column 1 (char 1)

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/Users/yy/miniconda3/bin/tira-cli", line 8, in <module>
    sys.exit(main())
  File "/Users/yy/miniconda3/lib/python3.10/site-packages/tira/tira_cli.py", line 531, in main
    return args.executable(**vars(args))
  File "/Users/yy/miniconda3/lib/python3.10/site-packages/tira/tira_cli.py", line 325, in code_submission_command
    client.submit_code(
  File "/Users/yy/miniconda3/lib/python3.10/site-packages/tira/tira_client.py", line 581, in submit_code
    upload = self.add_docker_software(
  File "/Users/yy/miniconda3/lib/python3.10/site-packages/tira/rest_api_client.py", line 321, in add_docker_software
    raise ValueError(msg)
ValueError: Upload of software failed with error 
<!doctype html>
<html lang="en">
<head>
  <title>Server Error (500)</title>
</head>
<body>
  <h1>Server Error (500)</h1><p></p>
</body>
</html>
 and response code 500.
2

Dear @yyLeaves,

thanks for reaching out and for participating in the Watermarking task, we are happy about this :slight_smile:

I believe this is the problem: Ensure To Upload Compatible Docker Images

The problem is that Docker Desktop (I assume you use Docker Desktop) builds since a recent update Docker images that are not compatible with the Docker registry of TIRA, but the Docker registry that TIRA uses does not recognize this during pushing the image, only when the image is pulled :slight_smile:

I pushed a new version of the TIRA client to PyPi, can you please run pip3 install --upgrade tira, and then follow the steps in this post. This should resolve the problem (it should mostly be disabling containerd in Docker Desktop).

Thanks in advance and please do not hesitate to contact us in case of other questions or problems.

Best regards,

Maik

3

Hi, thanks for the advices.

After a successful tira-cli code-submission, triggering a run on the spot-check dataset produces:

  • Time: 0

  • output/ directory exists but is empty (0 directories, 0 files)

  • STDOUT empty, STDERR empty (aside from the downstream evaluator’s ValueError: directory does not exist)

Could you spot the reason why the container isn’t being executed?

Thanks!