Success in debugging but failed in live leaderboard

Success in debugging but failed in live leaderboard ¶

By: sgupta1912 on Dec. 19, 2024, 4:23 a.m.

I tried submitting in live leaderboard after successfully submitting in debugging. But it failed, please help

Re: Success in debugging but failed in live leaderboard ¶

By: LindaSt on Dec. 19, 2024, 11:26 a.m.

Hi! The error log shows AttributeError: 'NoneType' object has no attribute 'shape'. You probably didn't run with the docker you thought you did :)...

2024-12-19T04:13:58.945000+00:00 /opt/conda/lib/python3.10/site-packages/torch/functional.py:504: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at /opt/conda/conda-bld/pytorch_1695392020201/work/aten/src/ATen/native/TensorShape.cpp:3526.)
2024-12-19T04:13:58.945000+00:00   return _VF.meshgrid(tensors, **kwargs)  # type: ignore[attr-defined]
2024-12-19T04:14:02.946000+00:00 Traceback (most recent call last):
2024-12-19T04:14:02.946000+00:00   File "/opt/app/inference.py", line 156, in <module>
2024-12-19T04:14:02.946000+00:00     raise SystemExit(run())
2024-12-19T04:14:02.946000+00:00   File "/opt/app/inference.py", line 37, in run
2024-12-19T04:14:02.946000+00:00     output_dict_lymphocytes, output_dict_monocytes, output_dict_inflammatory_cells = inference(image_path, mask_path, weight_root)
2024-12-19T04:14:02.946000+00:00   File "/opt/app/inference.py", line 84, in inference
2024-12-19T04:14:02.946000+00:00     annots = run_inference_dino(model, postprocessors, ts, mask, 416, 32, 8, 0.2, 480, {1:"lymphocytes",2:"monocytes",0:"N/A"}, device, {"N/A": 0, "monocytes": 0.2249391499719154, "lymphocytes": 0.24798367346938777}, 0.2, True)
2024-12-19T04:14:02.946000+00:00   File "/opt/app/utils_run.py", line 274, in run_inference_dino
2024-12-19T04:14:02.946000+00:00     ds = WSIObjectDetectionDataset(ts, transform, tile_size, pad_size, mask_img)
2024-12-19T04:14:02.946000+00:00   File "/opt/app/utils_run.py", line 61, in __init__
2024-12-19T04:14:02.946000+00:00     ydim, xdim = mask_img.shape
2024-12-19T04:14:02.946000+00:00 AttributeError: 'NoneType' object has no attribute 'shape'

Re: Success in debugging but failed in live leaderboard ¶

By: sgupta1912 on Dec. 19, 2024, 7:43 p.m.

Hey, I made only one successfull submission to debugging phase, and it says on live leaderboard phase you can only submit algorithms that is submitted successfully on debugging phase. I used this algo's https://grand-challenge.org/algorithms/monkeyv1/ latest version on both the phases. Their ids is also same, i.e. monkeyv1 (Image Version 1d93f177-0c20-425a-b9dc-a260380927b9 Model version 2bd3c5b6-2cf1-40b4-b36b-810fbc37d050 )

Re: Success in debugging but failed in live leaderboard ¶

By: LindaSt on Dec. 20, 2024, 11:05 a.m.

Hmm that's odd. I'm looking into it with the help of GC.

Re: Success in debugging but failed in live leaderboard ¶

By: sgupta1912 on Dec. 24, 2024, 2:07 a.m.

Hi @LindaST, did you get a chance to look at this?

Re: Success in debugging but failed in live leaderboard ¶

By: LindaSt on Dec. 24, 2024, 1:08 p.m.

They said it's likely an issue in your code. However, there are currently three cases in the validation set that have some issues. We were finally able to narrow it down and hope to post a fix in the upcoming days (it's the holidays here, so people are off thus it's taking a bit longer).