The reasoning of Docker submission guideline case cannot save the reasoning result, and an error is reported, indicating that the file saving path does not exist

When I was making docker, I referred to the case of docker submission guideline, but the final reasoning file could not be saved. Prompt: cannot open output file XXXXX. I also tried the official case, but also saved it incorrectly. I don't know what the problem is？

Did you mean that this error occured when you were running the docker container on your machine? Could you please provide the running command?

If you used the command in the submission guidelines (docker run ...), please confirm these points:

1. The container output path is the same as the code output path. It means that if your code output images to '/output', when you are running the container, the container path should be '/output' as well (no matter you specify it using the command or GUI).
2. The host output path on your machine should be writable.

We have tested our dummy submission and it works correctly. Please give us more information, for example the complete error log, if you have futher questions.

Thansk for your participation in KiPA22.

I'm testing the case given by the official：https://github.com/KiPA2022/kipa22/tree/main/DUMMY_SUBMISSION。I put the official case code under the /opt/code/ path of the machine, so I didn't modify and provide the dockerfile，And no other code has been modified. Then I created input and output folders in other paths of the machine and stored the data in the input, and then executed $docker run --rm --gpus all --runtime=nvidia --ipc =host -v /home/wangyi/workspace/input:/input/: RO -v /home/wangyi/workspace/output:/output/ first However, The final reasoning result is as follows： Predict test data 0.nii.gz ** ERROR （nifti_image_write_hdr_img2）:cannot open output file '/output/0.nii.gz'

The reason for this error is that the final reasoning result has not been saved, but logically, the input path has been read in and the data has been predicted, but the output path cannot be found here. I don't know why. I didn't change any code

I think this is a problem related to the host. Could you please confirm that your host machine has space left for the output files and your user account has the write access to '/home/wangyi/workspace/output'? Or could you try to run it on other machines?

You can see the container mount information by this command "dokcer inspect CONTAINER_ID".

Besides, I noticed that you have sent us a docker image as your first closed test submission. And it works successfully on our machine actually.

I did submit a docker for closed test. If it can run successfully in your place, I think it should be the problem of my own machine environment I will change the machine and test it again. Now that my first submission has run successfully, I hope you can return an email confirmation as soon as possible and provide a public test set submission. If I can't find the environmental problem in the end, I will still submit my final closed test docker in the way of generating docker for the first time to ensure that it is correct.

I have a problem. Since my first closed test was submitted successfully, I can get a chance to verify the public test set according to the rules. However, because I didn't read the rules carefully before, I didn't know that I could only submit the results three times, which led me to submit two more results on July 22. Can I directly delete the two results without returning them to me. Then I submit it again by myself. I don't know if this is OK

It's OK. You can create a submission on the open test as your 4th chance.

Thank you very much. At Aug. 5, 2022, 7:40 p.m. I submitted my fourth public verification test results

Hello, the organizer. The docker submitted for the first time has the problem of reasoning path on my own machine. A few days ago, you also provided a solution and explained that there is no problem in running on your machine. However, I did not solve this problem later, so my second docker still has the same problem as the first one. In order to ensure that it is correct on your machine, I hope you can give me timely feedback on the successful operation after the successful operation, and tell me whether the fifth public test set can be verified. Thank you very much.