![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F83267%2F4d0dc870b8879102e50463007ea302ab%2Fsubmission_scoring_error.png?generation=1669687168872049&alt=media\")
Hey,
\nTowards the end of my EDA : 📊 Initial EDA -- a first look at the data 🚀 I am outputting the predictions as follows:
\nsubmission = pd.DataFrame(data={'prediction_id': test_csv['prediction_id'], 'cancer': np.random.rand(test_csv.shape[0])})\nsubmission.to_csv('submission.csv', index=False)\nHowever, regardless of whether I output floats in the range of [0,1) as predictions, or all zeros, I am met with submission scoring error.
Could I please ask for help in figuring out what might be going wrong there?
\n
Am I doing something wrong or is there some issue with the machinery for scoring?
\nThank you for your help!
", "messages": [ { "id": 2047781, "postDate": "2022-11-29T03:30:11.060Z", "content": "Maybe, but I think it's because your submission.csv has duplicates for prediction_id.
", "rawMarkdown": "Maybe, but I think it's because your submission.csv has duplicates for prediction_id.", "votes": 7, "replies": [ { "id": 2047784, "postDate": "2022-11-29T03:32:40.687Z", "content": "Great suggestion, thank you!!!! will give this a go, that makes sense!
", "rawMarkdown": "Great suggestion, thank you!!!! will give this a go, that makes sense!", "votes": 1 }, { "id": 2047792, "postDate": "2022-11-29T03:42:23.403Z", "content": "That was it, thank you very much!!!!!!!!! 🙏
", "rawMarkdown": "That was it, thank you very much!!!!!!!!! 🙏", "votes": 2 } ] }, { "id": 2047727, "postDate": "2022-11-29T01:59:41.800Z", "content": "Hey,
\nTowards the end of my EDA : 📊 Initial EDA -- a first look at the data 🚀 I am outputting the predictions as follows:
\nsubmission = pd.DataFrame(data={'prediction_id': test_csv['prediction_id'], 'cancer': np.random.rand(test_csv.shape[0])})\nsubmission.to_csv('submission.csv', index=False)\nHowever, regardless of whether I output floats in the range of [0,1) as predictions, or all zeros, I am met with submission scoring error.
Could I please ask for help in figuring out what might be going wrong there?
\n
Am I doing something wrong or is there some issue with the machinery for scoring?
\nThank you for your help!
", "rawMarkdown": "Hey,\n\nTowards the end of my EDA : [📊 Initial EDA -- a first look at the data 🚀](https://www.kaggle.com/code/radek1/initial-eda-a-first-look-at-the-data/notebook?scriptVersionId=112383297) I am outputting the predictions as follows:\n\n```\nsubmission = pd.DataFrame(data={'prediction_id': test_csv['prediction_id'], 'cancer': np.random.rand(test_csv.shape[0])})\nsubmission.to_csv('submission.csv', index=False)\n```\n\nHowever, regardless of whether I output floats in the range of `[0,1)` as predictions, or all zeros, I am met with `submission scoring error`.\n\nCould I please ask for help in figuring out what might be going wrong there? \n\n\n\nAm I doing something wrong or is there some issue with the machinery for scoring?\n\nThank you for your help!", "votes": 8 }, { "id": 2054782, "postDate": "2022-12-04T13:04:40.187Z", "content": "you can simply prune duplicates by using max cancer predictions:
\ndf_preds = df_preds.groupby('prediction_id').max()
\nsee: https://www.kaggle.com/code/jirkaborovec/mammography-eda-loading-dicom
Hi, I have the same issue and I used the code above. I have unique prediction ID values and the csv is comma separated. However I am still getting the error message. However, when I submit the sample submission file it works. Would anyone be able to help me in this matter.
\nThanks & Best Regards
\nAMJS
Did you find any solution? My submission.csv looks like this, why is it still failing?
\nprediction_id cancer
\n10008_R 0.123
\n10008_L 0.123
these need to be comma seperated, they need to be a proper csv file
\nthe solution for me was removing duplicates, duplicated predictions for the same prediction_id
Do you drop duplicated prediction_id?
", "rawMarkdown": "Do you drop duplicated prediction_id?", "votes": 1, "replies": [ { "id": 2047785, "postDate": "2022-11-29T03:32:59.820Z", "content": "thank you @tomooinubushi, that makes sense 🙂 Trying it now!
", "rawMarkdown": "thank you @tomooinubushi, that makes sense 🙂 Trying it now!", "votes": 2 } ] }, { "id": 2047771, "postDate": "2022-11-29T03:13:10.363Z", "content": "Maybe you should name your submission file sample_submission.csv
", "rawMarkdown": "Maybe you should name your submission file sample_submission.csv", "replies": [ { "id": 2047782, "postDate": "2022-11-29T03:30:35.547Z", "content": "Thank you for the suggestion, @xiuqi0! That is an interesting thought, I think you might be right! Will give it a shot.
\nI looked at solutions from other competitions and think everything works there with the submission file being named submission.csv there. But might be not everything is set up properly here just yet 🙂
Trying now!
", "rawMarkdown": "Thank you for the suggestion, @xiuqi0! That is an interesting thought, I think you might be right! Will give it a shot.\n\nI looked at solutions from other competitions and think everything works there with the submission file being named `submission.csv` there. But might be not everything is set up properly here just yet 🙂\n\nTrying now! " } ] }, { "id": 2053295, "postDate": "2022-12-03T05:16:05.650Z", "content": "Wait, but combination prediction_id = \"patient_id + laterality\" is not unique, it corresponds to different image_id's.
\n How did you remove duplicated prediction_id's?
there is prediction_id that you can make sure is unique and all is well 🙂
Great, it works! thanks!!
", "rawMarkdown": "Great, it works! thanks!!" } ] } ], "comments": [ { "id": 2047781, "author_name": "YYama", "author_url": "", "post_date": "2022-11-29T03:30:11.060000", "content": "Maybe, but I think it's because your submission.csv has duplicates for prediction_id.
", "votes": 7, "replies": [ { "id": 2047784, "author_name": "Radek Osmulski", "author_url": "", "post_date": "2022-11-29T03:32:40.687000", "content": "Great suggestion, thank you!!!! will give this a go, that makes sense!
", "votes": 1, "replies": [] }, { "id": 2047792, "author_name": "Radek Osmulski", "author_url": "", "post_date": "2022-11-29T03:42:23.403000", "content": "That was it, thank you very much!!!!!!!!! 🙏
", "votes": 2, "replies": [] } ] }, { "id": 2054782, "author_name": "Jirka", "author_url": "", "post_date": "2022-12-04T13:04:40.187000", "content": "you can simply prune duplicates by using max cancer predictions:
\ndf_preds = df_preds.groupby('prediction_id').max()
\nsee: https://www.kaggle.com/code/jirkaborovec/mammography-eda-loading-dicom
Hi, I have the same issue and I used the code above. I have unique prediction ID values and the csv is comma separated. However I am still getting the error message. However, when I submit the sample submission file it works. Would anyone be able to help me in this matter.
\nThanks & Best Regards
\nAMJS
Did you find any solution? My submission.csv looks like this, why is it still failing?
\nprediction_id cancer
\n10008_R 0.123
\n10008_L 0.123
these need to be comma seperated, they need to be a proper csv file
\nthe solution for me was removing duplicates, duplicated predictions for the same prediction_id
Do you drop duplicated prediction_id?
", "votes": 1, "replies": [ { "id": 2047785, "author_name": "Radek Osmulski", "author_url": "", "post_date": "2022-11-29T03:32:59.820000", "content": "thank you @tomooinubushi, that makes sense 🙂 Trying it now!
", "votes": 2, "replies": [] } ] }, { "id": 2047771, "author_name": "XIUQI", "author_url": "", "post_date": "2022-11-29T03:13:10.363000", "content": "Maybe you should name your submission file sample_submission.csv
", "votes": 0, "replies": [ { "id": 2047782, "author_name": "Radek Osmulski", "author_url": "", "post_date": "2022-11-29T03:30:35.547000", "content": "Thank you for the suggestion, @xiuqi0! That is an interesting thought, I think you might be right! Will give it a shot.
\nI looked at solutions from other competitions and think everything works there with the submission file being named submission.csv there. But might be not everything is set up properly here just yet 🙂
Trying now!
", "votes": 0, "replies": [] } ] }, { "id": 2053295, "author_name": "Boris Polishchuk", "author_url": "", "post_date": "2022-12-03T05:16:05.650000", "content": "Wait, but combination prediction_id = \"patient_id + laterality\" is not unique, it corresponds to different image_id's.
\n How did you remove duplicated prediction_id's?
there is prediction_id that you can make sure is unique and all is well 🙂
Great, it works! thanks!!
", "votes": 0, "replies": [] } ] } ], "raw_markdown_by_id": { "2047781": "Maybe, but I think it's because your submission.csv has duplicates for prediction_id.", "2047727": "Hey,\n\nTowards the end of my EDA : [📊 Initial EDA -- a first look at the data 🚀](https://www.kaggle.com/code/radek1/initial-eda-a-first-look-at-the-data/notebook?scriptVersionId=112383297) I am outputting the predictions as follows:\n\n```\nsubmission = pd.DataFrame(data={'prediction_id': test_csv['prediction_id'], 'cancer': np.random.rand(test_csv.shape[0])})\nsubmission.to_csv('submission.csv', index=False)\n```\n\nHowever, regardless of whether I output floats in the range of `[0,1)` as predictions, or all zeros, I am met with `submission scoring error`.\n\nCould I please ask for help in figuring out what might be going wrong there? \n\n\n\nAm I doing something wrong or is there some issue with the machinery for scoring?\n\nThank you for your help!", "2054782": "you can simply prune duplicates by using max cancer predictions:\n`df_preds = df_preds.groupby('prediction_id').max()`\nsee: https://www.kaggle.com/code/jirkaborovec/mammography-eda-loading-dicom", "2050198": "Did you find any solution? My submission.csv looks like this, why is it still failing?\n\nprediction_id cancer\n10008_R 0.123\n10008_L 0.123", "2047780": "Do you drop duplicated prediction_id?", "2047771": "Maybe you should name your submission file sample_submission.csv", "2053295": "Wait, but combination prediction_id = \"patient_id + laterality\" is not unique, it corresponds to different image_id's.\n How did you remove duplicated prediction_id's? " } }