{
  "id": 183689,
  "title": "errors for submitting results",
  "url": "/competitions/rsna-str-pulmonary-embolism-detection/discussion/183689",
  "author_name": "David",
  "post_date": "2020-09-17T17:46:56.223000",
  "votes": 0,
  "comment_count": 10,
  "views": 0,
  "content": "<p>hi,  all</p>\n<p>I encountered an error in submitting the submission.csv for evaluation.</p>\n<ol>\n<li>I submit the sample_submission.csv to evaluate. Successful.</li>\n<li>I submit the submission.csv(attachments) to evaluate. Errors.</li>\n</ol>\n<p>I have checked the two files. The contents of both files are same.</p>\n<p>Could you guys be kind to help me to find the reason?</p>\n<p>thanks</p>",
  "messages": [
    {
      "id": 1014830,
      "postDate": "2020-09-17T18:08:15.197Z",
      "content": "<p>Remember you cannot see the real submission file.</p>\n<p>Best method to make sure your file is valid is to read in the private submission file and update it in place.</p>\n<p>Is your submission using pre-built data from the public test.csv and test images? It also needs to access the private test.csv/image files that can only be accessed in a committed notebook. The submission file you see is only from the public test data.</p>\n<p>-Rich</p>",
      "rawMarkdown": "Remember you cannot see the real submission file.\n\nBest method to make sure your file is valid is to read in the private submission file and update it in place.\n\nIs your submission using pre-built data from the public test.csv and test images? It also needs to access the private test.csv/image files that can only be accessed in a committed notebook. The submission file you see is only from the public test data.\n\n-Rich",
      "votes": 1,
      "replies": [
        {
          "id": 1058669,
          "postDate": "2020-10-24T04:56:29.333Z",
          "content": "<p>Hi, it is always so confusing… </p>\n<p>So, how to \"read in the private submission file\"? I am reading the file '/kaggle/input/rsna-str-pulmonary-embolism-detection/sample_submission.csv' in my notebook. How can I find that file you mention?</p>\n<p>The second one you mention is more confusing. I am need \"also\" access private test.csv/image files. Maybe I should use the that private submission file and then get the images from this private images?</p>\n<p>Can you elaborate more on that, please? I will appreciate.</p>\n<p>Thanks in advance! </p>",
          "rawMarkdown": "Hi, it is always so confusing... \n\nSo, how to \"read in the private submission file\"? I am reading the file '/kaggle/input/rsna-str-pulmonary-embolism-detection/sample_submission.csv' in my notebook. How can I find that file you mention?\n\nThe second one you mention is more confusing. I am need \"also\" access private test.csv/image files. Maybe I should use the that private submission file and then get the images from this private images?\n\nCan you elaborate more on that, please? I will appreciate.\n\nThanks in advance! ",
          "replies": [
            {
              "id": 1058908,
              "postDate": "2020-10-24T12:09:51.600Z",
              "content": "<p>When  you run your code interactively or in a normal commit, you are running on the public test csv/data.</p>\n<p>When you then \"submit to competition\" that notebook (under \"Submit Predictions\") it will automatically use the hidden test csv/data.</p>\n<p>Your code runs exactly the same, but it is running against a new test.csv, sample_submission.csv and image data.</p>",
              "rawMarkdown": "When  you run your code interactively or in a normal commit, you are running on the public test csv/data.\n\nWhen you then \"submit to competition\" that notebook (under \"Submit Predictions\") it will automatically use the hidden test csv/data.\n\nYour code runs exactly the same, but it is running against a new test.csv, sample_submission.csv and image data.",
              "votes": 1
            },
            {
              "id": 1061201,
              "postDate": "2020-10-26T19:53:16.437Z",
              "content": "<p>Thanks a lot. It is clear.</p>",
              "rawMarkdown": "Thanks a lot. It is clear."
            }
          ]
        }
      ]
    },
    {
      "id": 1021534,
      "postDate": "2020-09-22T00:29:58.580Z",
      "content": "<p>This is a inference-only competition. That means when you submit a kernel to this competition, it will be re-evaluated (not-training) on  a new unseen test set. So when you are putting *sample_submission.csv* it is actually submitting the files with the file names from the unseen test set and it works. </p>\n\n<p>It might be the case that, somehow, <em>submission.csv</em> is somehow hardcoded and it is not getting those filenames from the hidden test_set. </p>\n\n<p>I encountered the same problem  and to solve that and just for the demonstration purpose, I made this notebook <a href=\"https://www.kaggle.com/redwankarimsony/rsna-str-pulmonary-embolism-dummy-sub\">RSNA-STR Pulmonary Embolism [Dummy Sub]</a>.</p>\n\n<p>You can have a look to understand what's going.. </p>\n\n<p>Have a good day!!!</p>",
      "rawMarkdown": "This is a inference-only competition. That means when you submit a kernel to this competition, it will be re-evaluated (not-training) on  a new unseen test set. So when you are putting *sample_submission.csv* it is actually submitting the files with the file names from the unseen test set and it works. \n\nIt might be the case that, somehow, *submission.csv* is somehow hardcoded and it is not getting those filenames from the hidden test_set. \n\nI encountered the same problem  and to solve that and just for the demonstration purpose, I made this notebook [RSNA-STR Pulmonary Embolism [Dummy Sub]](https://www.kaggle.com/redwankarimsony/rsna-str-pulmonary-embolism-dummy-sub).\n\nYou can have a look to understand what's going.. \n\nHave a good day!!!\n"
    },
    {
      "id": 1015329,
      "postDate": "2020-09-18T05:53:51.363Z",
      "content": "<p><a href=\"https://www.kaggle.com/richardepstein\" target=\"_blank\">@richardepstein</a></p>\n<p>hi,</p>\n<p>It also needs to access the private test.csv/image files that can only be accessed in a committed notebook.</p>\n<p>The test.csv/image in notebook only has 146853 rows, which is the same as the test.csv in the public dataset.<br>\nI know that my submission file is only from the public test data. But the sample_submission file is also only from the public test data. Why we can just submit the sample_submission file and obtain a score?</p>\n<p>In addition,  I didn't see any private test.csv in notebook? how to obtain this?</p>\n<p>Looking forward to your reply<br>\nThanks,<br>\nzw</p>",
      "rawMarkdown": "@richardepstein\n\nhi,\n\nIt also needs to access the private test.csv/image files that can only be accessed in a committed notebook.\n\nThe test.csv/image in notebook only has 146853 rows, which is the same as the test.csv in the public dataset.\nI know that my submission file is only from the public test data. But the sample_submission file is also only from the public test data. Why we can just submit the sample_submission file and obtain a score?\n\nIn addition,  I didn't see any private test.csv in notebook? how to obtain this?\n\nLooking forward to your reply\nThanks,\nzw",
      "replies": [
        {
          "id": 1021548,
          "postDate": "2020-09-22T01:21:06.650Z",
          "content": "<p>Not sure if you are still having the issue:</p>\n<p>When you run a notebook interactively, it shows the \"visible\" test.csv and sample_submission.csv\" and test images.</p>\n<p>When you submit a notebook for scoring, it runs against the \"hidden\" test.csv and sample_submission.csv\" and test images. The names are the same. You can never see the \"hidden\" files.</p>\n<p>Submitting for scoring is a multi-step process:</p>\n<ol>\n<li><p>Save/commit your notebook and run it. It will run against the \"visible\" data. When finished, you can see the results, but these are \"visible\" results. It needs to complete successfully (and probably produce a \"submission.csv\" file).</p></li>\n<li><p>Under \"Submit Predictions\" it will list your notebook. You can then Submit your notebook to the competition.</p></li>\n<li><p>Your notebook will run (Can watch to see when if finishes under \"My Submissions\"). It is running against the \"hidden\" data. You cannot see any output. The error log is basically \"I failed\" without details. You have to time the notebook yourself. The log doesn't provide how long it took.</p></li>\n<li><p>Notice that you have to run your notebook successfully before you can submit it. I have code in my notebook that checks the \"test.csv\" file length.  If I am running against \"visible\" data, I skip most of my processing (like only do inference on 100 records). Speeds up the process (assuming you are confident your code works):</p></li>\n</ol>\n<p>df = pd.read_csv(mypath+'/test.csv')<br>\nif len(df.index) == 146853:         #running against \"visible\" test data<br>\n  code to skip most processing</p>\n<ol>\n<li><p>The \"hidden\" test data includes all the \"visible\" test data also. So you can preprocess/inference the \"visible\" test data, copy a submission file into a dataset, and then in your Committed Notebook only process/inference the rest of the \"hidden\" test data and combine it with your \"visible\" submission file. This will cut your processing by 40%.</p></li>\n<li><p>For faster testing, do the inference just on the \"visible\" test data using a TPU. Copy that submission file into a dataset and commit a notebook to update the \"hidden\" sample_submission.csv with your submission file. You are only updating 40% of the \"hidden\" submission, but you can see if you are improving on default values. When you think things are going well, run step 5 (above) with the full \"visible\" + \"hidden\" data to see your real score.</p></li>\n<li><p>As usual in these competitions, the Leaderboard only shows you a score based on a subset of your data. I don't think we know if the Leaderboard score division of data is related to the \"visible\" and \"hidden\" test datasets.</p></li>\n</ol>\n<p>Good luck,</p>\n<p>-Rich</p>",
          "rawMarkdown": "Not sure if you are still having the issue:\n\nWhen you run a notebook interactively, it shows the \"visible\" test.csv and sample_submission.csv\" and test images.\n\nWhen you submit a notebook for scoring, it runs against the \"hidden\" test.csv and sample_submission.csv\" and test images. The names are the same. You can never see the \"hidden\" files.\n\nSubmitting for scoring is a multi-step process:\n\n1. Save/commit your notebook and run it. It will run against the \"visible\" data. When finished, you can see the results, but these are \"visible\" results. It needs to complete successfully (and probably produce a \"submission.csv\" file).\n\n2. Under \"Submit Predictions\" it will list your notebook. You can then Submit your notebook to the competition.\n\n3. Your notebook will run (Can watch to see when if finishes under \"My Submissions\"). It is running against the \"hidden\" data. You cannot see any output. The error log is basically \"I failed\" without details. You have to time the notebook yourself. The log doesn't provide how long it took.\n\n4. Notice that you have to run your notebook successfully before you can submit it. I have code in my notebook that checks the \"test.csv\" file length.  If I am running against \"visible\" data, I skip most of my processing (like only do inference on 100 records). Speeds up the process (assuming you are confident your code works):\n\ndf = pd.read_csv(mypath+'/test.csv')\nif len(df.index) == 146853:         #running against \"visible\" test data\n  code to skip most processing\n\n5. The \"hidden\" test data includes all the \"visible\" test data also. So you can preprocess/inference the \"visible\" test data, copy a submission file into a dataset, and then in your Committed Notebook only process/inference the rest of the \"hidden\" test data and combine it with your \"visible\" submission file. This will cut your processing by 40%.\n\n6. For faster testing, do the inference just on the \"visible\" test data using a TPU. Copy that submission file into a dataset and commit a notebook to update the \"hidden\" sample_submission.csv with your submission file. You are only updating 40% of the \"hidden\" submission, but you can see if you are improving on default values. When you think things are going well, run step 5 (above) with the full \"visible\" + \"hidden\" data to see your real score.\n\n7. As usual in these competitions, the Leaderboard only shows you a score based on a subset of your data. I don't think we know if the Leaderboard score division of data is related to the \"visible\" and \"hidden\" test datasets.\n\nGood luck,\n\n-Rich"
        }
      ]
    },
    {
      "id": 1015309,
      "postDate": "2020-09-18T05:38:16.987Z",
      "content": "<p><a href=\"https://www.kaggle.com/richardepstein\" target=\"_blank\">@richardepstein</a> </p>\n<p>hi, Richard</p>\n<p>I met the same problem.</p>\n<p>I obtain the scores of each image. I try to replace the scores in sample_submission.csv by our scores.<br>\nThen, I upload the results to notebook and submit to evaluate.<br>\nFinally, it returns <em>Submission Scoring Error.</em></p>\n<p>I have checked the results carefully. And all of the scores are between 0 and 1.<br>\nAnd my submission.csv has 152703 rows, which is the same  as sample_submission.csv.</p>\n<p>I'm very confused about this problem.</p>\n<p>Do you have any ideas to debug?</p>\n<p>thanks</p>",
      "rawMarkdown": "@richardepstein \n\nhi, Richard\n\nI met the same problem.\n\nI obtain the scores of each image. I try to replace the scores in sample_submission.csv by our scores.\nThen, I upload the results to notebook and submit to evaluate.\nFinally, it returns *Submission Scoring Error.*\n\n\nI have checked the results carefully. And all of the scores are between 0 and 1.\nAnd my submission.csv has 152703 rows, which is the same  as sample_submission.csv.\n\nI'm very confused about this problem.\n\nDo you have any ideas to debug?\n\nthanks",
      "replies": [
        {
          "id": 1020379,
          "postDate": "2020-09-21T06:11:14.247Z",
          "content": "<p>Hey, <a href=\"https://www.kaggle.com/qiuzhongwei\" target=\"_blank\">@qiuzhongwei</a> did you figured it out? I am getting the same error and unable to resolve it.</p>",
          "rawMarkdown": "Hey, @qiuzhongwei did you figured it out? I am getting the same error and unable to resolve it."
        }
      ]
    },
    {
      "id": 1014886,
      "postDate": "2020-09-17T19:06:37.837Z",
      "content": "<p>In addition to following Richard's helpful guidance/reply, our <a href=\"https://www.kaggle.com/code-competition-debugging\" target=\"_blank\">Code Debugging doc</a> may help you explore possible sources of errors.</p>",
      "rawMarkdown": "In addition to following Richard's helpful guidance/reply, our [Code Debugging doc](https://www.kaggle.com/code-competition-debugging) may help you explore possible sources of errors."
    },
    {
      "id": 1014793,
      "postDate": "2020-09-17T17:46:56.223Z",
      "content": "<p>hi,  all</p>\n<p>I encountered an error in submitting the submission.csv for evaluation.</p>\n<ol>\n<li>I submit the sample_submission.csv to evaluate. Successful.</li>\n<li>I submit the submission.csv(attachments) to evaluate. Errors.</li>\n</ol>\n<p>I have checked the two files. The contents of both files are same.</p>\n<p>Could you guys be kind to help me to find the reason?</p>\n<p>thanks</p>",
      "rawMarkdown": "hi,  all\n\nI encountered an error in submitting the submission.csv for evaluation.\n\n1. I submit the sample_submission.csv to evaluate. Successful.\n2. I submit the submission.csv(attachments) to evaluate. Errors.\n\nI have checked the two files. The contents of both files are same.\n\nCould you guys be kind to help me to find the reason?\n\nthanks\n\n\n"
    }
  ],
  "comments": [
    {
      "id": 1014830,
      "author_name": "quadcore/Richard Epstein",
      "author_url": "",
      "post_date": "2020-09-17T18:08:15.197000",
      "content": "<p>Remember you cannot see the real submission file.</p>\n<p>Best method to make sure your file is valid is to read in the private submission file and update it in place.</p>\n<p>Is your submission using pre-built data from the public test.csv and test images? It also needs to access the private test.csv/image files that can only be accessed in a committed notebook. The submission file you see is only from the public test data.</p>\n<p>-Rich</p>",
      "votes": 1,
      "replies": [
        {
          "id": 1058669,
          "author_name": "Juliano Batistim",
          "author_url": "",
          "post_date": "2020-10-24T04:56:29.333000",
          "content": "<p>Hi, it is always so confusing… </p>\n<p>So, how to \"read in the private submission file\"? I am reading the file '/kaggle/input/rsna-str-pulmonary-embolism-detection/sample_submission.csv' in my notebook. How can I find that file you mention?</p>\n<p>The second one you mention is more confusing. I am need \"also\" access private test.csv/image files. Maybe I should use the that private submission file and then get the images from this private images?</p>\n<p>Can you elaborate more on that, please? I will appreciate.</p>\n<p>Thanks in advance! </p>",
          "votes": 0,
          "replies": [
            {
              "id": 1058908,
              "author_name": "quadcore/Richard Epstein",
              "author_url": "",
              "post_date": "2020-10-24T12:09:51.600000",
              "content": "<p>When  you run your code interactively or in a normal commit, you are running on the public test csv/data.</p>\n<p>When you then \"submit to competition\" that notebook (under \"Submit Predictions\") it will automatically use the hidden test csv/data.</p>\n<p>Your code runs exactly the same, but it is running against a new test.csv, sample_submission.csv and image data.</p>",
              "votes": 1,
              "replies": []
            },
            {
              "id": 1061201,
              "author_name": "Juliano Batistim",
              "author_url": "",
              "post_date": "2020-10-26T19:53:16.437000",
              "content": "<p>Thanks a lot. It is clear.</p>",
              "votes": 0,
              "replies": []
            }
          ]
        }
      ]
    },
    {
      "id": 1021534,
      "author_name": "Redwan Sony",
      "author_url": "",
      "post_date": "2020-09-22T00:29:58.580000",
      "content": "<p>This is a inference-only competition. That means when you submit a kernel to this competition, it will be re-evaluated (not-training) on  a new unseen test set. So when you are putting *sample_submission.csv* it is actually submitting the files with the file names from the unseen test set and it works. </p>\n\n<p>It might be the case that, somehow, <em>submission.csv</em> is somehow hardcoded and it is not getting those filenames from the hidden test_set. </p>\n\n<p>I encountered the same problem  and to solve that and just for the demonstration purpose, I made this notebook <a href=\"https://www.kaggle.com/redwankarimsony/rsna-str-pulmonary-embolism-dummy-sub\">RSNA-STR Pulmonary Embolism [Dummy Sub]</a>.</p>\n\n<p>You can have a look to understand what's going.. </p>\n\n<p>Have a good day!!!</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 1015329,
      "author_name": "qzw",
      "author_url": "",
      "post_date": "2020-09-18T05:53:51.363000",
      "content": "<p><a href=\"https://www.kaggle.com/richardepstein\" target=\"_blank\">@richardepstein</a></p>\n<p>hi,</p>\n<p>It also needs to access the private test.csv/image files that can only be accessed in a committed notebook.</p>\n<p>The test.csv/image in notebook only has 146853 rows, which is the same as the test.csv in the public dataset.<br>\nI know that my submission file is only from the public test data. But the sample_submission file is also only from the public test data. Why we can just submit the sample_submission file and obtain a score?</p>\n<p>In addition,  I didn't see any private test.csv in notebook? how to obtain this?</p>\n<p>Looking forward to your reply<br>\nThanks,<br>\nzw</p>",
      "votes": 0,
      "replies": [
        {
          "id": 1021548,
          "author_name": "quadcore/Richard Epstein",
          "author_url": "",
          "post_date": "2020-09-22T01:21:06.650000",
          "content": "<p>Not sure if you are still having the issue:</p>\n<p>When you run a notebook interactively, it shows the \"visible\" test.csv and sample_submission.csv\" and test images.</p>\n<p>When you submit a notebook for scoring, it runs against the \"hidden\" test.csv and sample_submission.csv\" and test images. The names are the same. You can never see the \"hidden\" files.</p>\n<p>Submitting for scoring is a multi-step process:</p>\n<ol>\n<li><p>Save/commit your notebook and run it. It will run against the \"visible\" data. When finished, you can see the results, but these are \"visible\" results. It needs to complete successfully (and probably produce a \"submission.csv\" file).</p></li>\n<li><p>Under \"Submit Predictions\" it will list your notebook. You can then Submit your notebook to the competition.</p></li>\n<li><p>Your notebook will run (Can watch to see when if finishes under \"My Submissions\"). It is running against the \"hidden\" data. You cannot see any output. The error log is basically \"I failed\" without details. You have to time the notebook yourself. The log doesn't provide how long it took.</p></li>\n<li><p>Notice that you have to run your notebook successfully before you can submit it. I have code in my notebook that checks the \"test.csv\" file length.  If I am running against \"visible\" data, I skip most of my processing (like only do inference on 100 records). Speeds up the process (assuming you are confident your code works):</p></li>\n</ol>\n<p>df = pd.read_csv(mypath+'/test.csv')<br>\nif len(df.index) == 146853:         #running against \"visible\" test data<br>\n  code to skip most processing</p>\n<ol>\n<li><p>The \"hidden\" test data includes all the \"visible\" test data also. So you can preprocess/inference the \"visible\" test data, copy a submission file into a dataset, and then in your Committed Notebook only process/inference the rest of the \"hidden\" test data and combine it with your \"visible\" submission file. This will cut your processing by 40%.</p></li>\n<li><p>For faster testing, do the inference just on the \"visible\" test data using a TPU. Copy that submission file into a dataset and commit a notebook to update the \"hidden\" sample_submission.csv with your submission file. You are only updating 40% of the \"hidden\" submission, but you can see if you are improving on default values. When you think things are going well, run step 5 (above) with the full \"visible\" + \"hidden\" data to see your real score.</p></li>\n<li><p>As usual in these competitions, the Leaderboard only shows you a score based on a subset of your data. I don't think we know if the Leaderboard score division of data is related to the \"visible\" and \"hidden\" test datasets.</p></li>\n</ol>\n<p>Good luck,</p>\n<p>-Rich</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 1015309,
      "author_name": "qzw",
      "author_url": "",
      "post_date": "2020-09-18T05:38:16.987000",
      "content": "<p><a href=\"https://www.kaggle.com/richardepstein\" target=\"_blank\">@richardepstein</a> </p>\n<p>hi, Richard</p>\n<p>I met the same problem.</p>\n<p>I obtain the scores of each image. I try to replace the scores in sample_submission.csv by our scores.<br>\nThen, I upload the results to notebook and submit to evaluate.<br>\nFinally, it returns <em>Submission Scoring Error.</em></p>\n<p>I have checked the results carefully. And all of the scores are between 0 and 1.<br>\nAnd my submission.csv has 152703 rows, which is the same  as sample_submission.csv.</p>\n<p>I'm very confused about this problem.</p>\n<p>Do you have any ideas to debug?</p>\n<p>thanks</p>",
      "votes": 0,
      "replies": [
        {
          "id": 1020379,
          "author_name": "SumanSudhir",
          "author_url": "",
          "post_date": "2020-09-21T06:11:14.247000",
          "content": "<p>Hey, <a href=\"https://www.kaggle.com/qiuzhongwei\" target=\"_blank\">@qiuzhongwei</a> did you figured it out? I am getting the same error and unable to resolve it.</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 1014886,
      "author_name": "Julia Elliott",
      "author_url": "",
      "post_date": "2020-09-17T19:06:37.837000",
      "content": "<p>In addition to following Richard's helpful guidance/reply, our <a href=\"https://www.kaggle.com/code-competition-debugging\" target=\"_blank\">Code Debugging doc</a> may help you explore possible sources of errors.</p>",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1014830": "Remember you cannot see the real submission file.\n\nBest method to make sure your file is valid is to read in the private submission file and update it in place.\n\nIs your submission using pre-built data from the public test.csv and test images? It also needs to access the private test.csv/image files that can only be accessed in a committed notebook. The submission file you see is only from the public test data.\n\n-Rich",
    "1021534": "This is a inference-only competition. That means when you submit a kernel to this competition, it will be re-evaluated (not-training) on  a new unseen test set. So when you are putting *sample_submission.csv* it is actually submitting the files with the file names from the unseen test set and it works. \n\nIt might be the case that, somehow, *submission.csv* is somehow hardcoded and it is not getting those filenames from the hidden test_set. \n\nI encountered the same problem  and to solve that and just for the demonstration purpose, I made this notebook [RSNA-STR Pulmonary Embolism [Dummy Sub]](https://www.kaggle.com/redwankarimsony/rsna-str-pulmonary-embolism-dummy-sub).\n\nYou can have a look to understand what's going.. \n\nHave a good day!!!\n",
    "1015329": "@richardepstein\n\nhi,\n\nIt also needs to access the private test.csv/image files that can only be accessed in a committed notebook.\n\nThe test.csv/image in notebook only has 146853 rows, which is the same as the test.csv in the public dataset.\nI know that my submission file is only from the public test data. But the sample_submission file is also only from the public test data. Why we can just submit the sample_submission file and obtain a score?\n\nIn addition,  I didn't see any private test.csv in notebook? how to obtain this?\n\nLooking forward to your reply\nThanks,\nzw",
    "1015309": "@richardepstein \n\nhi, Richard\n\nI met the same problem.\n\nI obtain the scores of each image. I try to replace the scores in sample_submission.csv by our scores.\nThen, I upload the results to notebook and submit to evaluate.\nFinally, it returns *Submission Scoring Error.*\n\n\nI have checked the results carefully. And all of the scores are between 0 and 1.\nAnd my submission.csv has 152703 rows, which is the same  as sample_submission.csv.\n\nI'm very confused about this problem.\n\nDo you have any ideas to debug?\n\nthanks",
    "1014886": "In addition to following Richard's helpful guidance/reply, our [Code Debugging doc](https://www.kaggle.com/code-competition-debugging) may help you explore possible sources of errors.",
    "1014793": "hi,  all\n\nI encountered an error in submitting the submission.csv for evaluation.\n\n1. I submit the sample_submission.csv to evaluate. Successful.\n2. I submit the submission.csv(attachments) to evaluate. Errors.\n\nI have checked the two files. The contents of both files are same.\n\nCould you guys be kind to help me to find the reason?\n\nthanks\n\n\n"
  }
}