I give up, I am on the verge of abandoning this competition (MOA), I have been lost for dozens of hours waiting for the public result and receiving the error ( Notebook Exceeded Allowed Compute, Notebook Timeout), in all cases. I obtain the private result using the GPU that Kaggle provides me in less than 50 minutes , when I try to obtain the public result, it takes more than 9 hours to tell me ( Notebook Exceeded Allowed Compute, Notebook Timeout). I don't think that, as much as I change the files with which the notebook is tested, there is a fully justified reason for this to happen. Put in the contest rules the maximum execution time of the notebook both using GPU and using only CPU. I am partially disappointed, the positive part is what I have learned in this challenge.
\nI apologize for my English.
I give up, I am on the verge of abandoning this competition (MOA), I have been lost for dozens of hours waiting for the public result and receiving the error ( Notebook Exceeded Allowed Compute, Notebook Timeout), in all cases. I obtain the private result using the GPU that Kaggle provides me in less than 50 minutes , when I try to obtain the public result, it takes more than 9 hours to tell me ( Notebook Exceeded Allowed Compute, Notebook Timeout). I don't think that, as much as I change the files with which the notebook is tested, there is a fully justified reason for this to happen. Put in the contest rules the maximum execution time of the notebook both using GPU and using only CPU. I am partially disappointed, the positive part is what I have learned in this challenge.
\nI apologize for my English.
You have posted it in the wrong competition's discussion forum, you should have posted it here.
\nThe reason you're getting this error is because the test data is 4 times the train data. After you submit, they re-run your notebook on a different test set which is 4x the train data. So it takes 4 times longer to run your notebook. It is mentioned here.
", "rawMarkdown": "You have posted it in the wrong competition's discussion forum, you should have posted it [here](https://www.kaggle.com/c/lish-moa/discussion).\n\nThe reason you're getting this error is because the test data is 4 times the train data. After you submit, they re-run your notebook on a different test set which is 4x the train data. So it takes 4 times longer to run your notebook. It is mentioned [here](https://www.kaggle.com/c/lish-moa/data).", "votes": null } ], "comments": [ { "id": 1057396, "author_name": "rude009", "author_url": "", "post_date": "10/22/2020 15:54:35", "content": "You have posted it in the wrong competition's discussion forum, you should have posted it here.
\nThe reason you're getting this error is because the test data is 4 times the train data. After you submit, they re-run your notebook on a different test set which is 4x the train data. So it takes 4 times longer to run your notebook. It is mentioned here.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "1057047": "I give up, I am on the verge of abandoning this competition (MOA), I have been lost for dozens of hours waiting for the public result and receiving the error ( Notebook Exceeded Allowed Compute, Notebook Timeout), in all cases. I obtain the private result using the GPU that Kaggle provides me in less than 50 minutes , when I try to obtain the public result, it takes more than 9 hours to tell me ( Notebook Exceeded Allowed Compute, Notebook Timeout). I don't think that, as much as I change the files with which the notebook is tested, there is a fully justified reason for this to happen. Put in the contest rules the maximum execution time of the notebook both using GPU and using only CPU. I am partially disappointed, the positive part is what I have learned in this challenge.\nI apologize for my English.", "1057396": "You have posted it in the wrong competition's discussion forum, you should have posted it [here](https://www.kaggle.com/c/lish-moa/discussion).\n\nThe reason you're getting this error is because the test data is 4 times the train data. After you submit, they re-run your notebook on a different test set which is 4x the train data. So it takes 4 times longer to run your notebook. It is mentioned [here](https://www.kaggle.com/c/lish-moa/data)." }, "source": "meta" }