{
  "id": 152066,
  "title": "Problems with submitting the submission file",
  "url": "/competitions/global-wheat-detection/discussion/152066",
  "author_name": "",
  "post_date": "2020-05-18T09:08:37.357294300Z",
  "votes": 1,
  "comment_count": 6,
  "views": 0,
  "content": "<p>I'm keep trying to submit a submission file and no matter what I do i'm getting submission errors.\nThe only option that did worked for me was forking the \"global-wheat-detection-first-submission\" kernel. \nI then tried to create a kernel that use the submission file from that kernel and save it as csv file. i double checked with pd.equals that the file is identical and i'm still getting submission error.</p>\n\n<p>I even tried to create a kernel where i manually write the values from that kernel submission file to a new dataframe and save it as cvs, once again I double checked that my new submission.csv file is identical to the original one from the started kernel.</p>\n\n<p>It's frustrating that there is no explanation of why the submission is failing, any idea how to proceed?</p>",
  "messages": [
    {
      "id": "852241",
      "postDate": "05/18/2020 09:08:37",
      "content": "<p>I'm keep trying to submit a submission file and no matter what I do i'm getting submission errors.\nThe only option that did worked for me was forking the \"global-wheat-detection-first-submission\" kernel. \nI then tried to create a kernel that use the submission file from that kernel and save it as csv file. i double checked with pd.equals that the file is identical and i'm still getting submission error.</p>\n\n<p>I even tried to create a kernel where i manually write the values from that kernel submission file to a new dataframe and save it as cvs, once again I double checked that my new submission.csv file is identical to the original one from the started kernel.</p>\n\n<p>It's frustrating that there is no explanation of why the submission is failing, any idea how to proceed?</p>",
      "rawMarkdown": "I'm keep trying to submit a submission file and no matter what I do i'm getting submission errors.\nThe only option that did worked for me was forking the \"global-wheat-detection-first-submission\" kernel. \nI then tried to create a kernel that use the submission file from that kernel and save it as csv file. i double checked with pd.equals that the file is identical and i'm still getting submission error.\n\nI even tried to create a kernel where i manually write the values from that kernel submission file to a new dataframe and save it as cvs, once again I double checked that my new submission.csv file is identical to the original one from the started kernel.\n\nIt's frustrating that there is no explanation of why the submission is failing, any idea how to proceed?",
      "votes": null
    },
    {
      "id": "938578",
      "postDate": "07/21/2020 15:59:22",
      "content": "<p>I had the same issue. First make sure that the submission file is called submission.csv, then make sure that it has the same format as in the sample, it means that it has 2 columns where the first is the file name without jpg and then the string is Accuracy x y width heigth. When there are many objects in the same image, continue the object in the same line, so one image one row.</p>\n\n<p>My mistake was that I created a file only for submission. Remember that there are hidden images, not only those 10 images that you see in the test folder. So your prediction code needs to be in the same kernel, so the system can generate the code for the rest of the hidden images.</p>\n\n<p>Also disable internet as it is not allowed to use internet on the notebook that is submitting. If you need a package, crate a dataset or create another notebook and download the packages there, you can use !pip download \"package\" and it will download the wheel files that you need to install. You will need !pip install \"package-name\" \"your-path\"</p>\n\n<p>I hope it helps.</p>",
      "rawMarkdown": "I had the same issue. First make sure that the submission file is called submission.csv, then make sure that it has the same format as in the sample, it means that it has 2 columns where the first is the file name without jpg and then the string is Accuracy x y width heigth. When there are many objects in the same image, continue the object in the same line, so one image one row.\n\nMy mistake was that I created a file only for submission. Remember that there are hidden images, not only those 10 images that you see in the test folder. So your prediction code needs to be in the same kernel, so the system can generate the code for the rest of the hidden images.\n\nAlso disable internet as it is not allowed to use internet on the notebook that is submitting. If you need a package, crate a dataset or create another notebook and download the packages there, you can use !pip download \"package\" and it will download the wheel files that you need to install. You will need !pip install \"package-name\" \"your-path\"\n\nI hope it helps.",
      "votes": null
    },
    {
      "id": "956079",
      "postDate": "08/03/2020 08:17:03",
      "content": "<p>Hello, you said \"Also disable internet as it is not allowed to use internet on the notebook that is submitting\", I am wondering how possible we could connect to Kaggle if without Internet, we are using laptop, not cell phone.</p>",
      "rawMarkdown": "Hello, you said \"Also disable internet as it is not allowed to use internet on the notebook that is submitting\", I am wondering how possible we could connect to Kaggle if without Internet, we are using laptop, not cell phone.",
      "votes": null
    },
    {
      "id": "956086",
      "postDate": "08/03/2020 08:22:09",
      "content": "<p>No, it does not mean internet in your laptop. It means internet in the Kaggle Notebook. This is done to avoid adding any script downloaded. </p>\n\n<p>This link explains how to enable internet.</p>\n\n<p><a href=\"https://www.kaggle.com/product-feedback/63544\">https://www.kaggle.com/product-feedback/63544</a></p>",
      "rawMarkdown": "No, it does not mean internet in your laptop. It means internet in the Kaggle Notebook. This is done to avoid adding any script downloaded. \n\nThis link explains how to enable internet.\n\nhttps://www.kaggle.com/product-feedback/63544",
      "votes": null
    },
    {
      "id": "956115",
      "postDate": "08/03/2020 08:49:39",
      "content": "<p>Thank you, it looks like by default it is OFF. BTW, where can I find a submit button in the kernel? on the top I can only see \"share\" and \"save version\". Thanks.</p>",
      "rawMarkdown": "Thank you, it looks like by default it is OFF. BTW, where can I find a submit button in the kernel? on the top I can only see \"share\" and \"save version\". Thanks.",
      "votes": null
    },
    {
      "id": "956177",
      "postDate": "08/03/2020 10:00:33",
      "content": "<p>Yes, it is off by default. Use the save version button. Then save and run. This will commit and you will be able to see the output, then you can submit.</p>",
      "rawMarkdown": "Yes, it is off by default. Use the save version button. Then save and run. This will commit and you will be able to see the output, then you can submit.",
      "votes": null
    },
    {
      "id": "956494",
      "postDate": "08/03/2020 14:41:50",
      "content": "<p>Thank you!</p>",
      "rawMarkdown": "Thank you!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 938578,
      "author_name": "luisvivas",
      "author_url": "",
      "post_date": "07/21/2020 15:59:22",
      "content": "<p>I had the same issue. First make sure that the submission file is called submission.csv, then make sure that it has the same format as in the sample, it means that it has 2 columns where the first is the file name without jpg and then the string is Accuracy x y width heigth. When there are many objects in the same image, continue the object in the same line, so one image one row.</p>\n\n<p>My mistake was that I created a file only for submission. Remember that there are hidden images, not only those 10 images that you see in the test folder. So your prediction code needs to be in the same kernel, so the system can generate the code for the rest of the hidden images.</p>\n\n<p>Also disable internet as it is not allowed to use internet on the notebook that is submitting. If you need a package, crate a dataset or create another notebook and download the packages there, you can use !pip download \"package\" and it will download the wheel files that you need to install. You will need !pip install \"package-name\" \"your-path\"</p>\n\n<p>I hope it helps.</p>",
      "votes": null,
      "replies": [
        {
          "id": 956079,
          "author_name": "chunlin9",
          "author_url": "",
          "post_date": "08/03/2020 08:17:03",
          "content": "<p>Hello, you said \"Also disable internet as it is not allowed to use internet on the notebook that is submitting\", I am wondering how possible we could connect to Kaggle if without Internet, we are using laptop, not cell phone.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 956086,
          "author_name": "luisvivas",
          "author_url": "",
          "post_date": "08/03/2020 08:22:09",
          "content": "<p>No, it does not mean internet in your laptop. It means internet in the Kaggle Notebook. This is done to avoid adding any script downloaded. </p>\n\n<p>This link explains how to enable internet.</p>\n\n<p><a href=\"https://www.kaggle.com/product-feedback/63544\">https://www.kaggle.com/product-feedback/63544</a></p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 956115,
          "author_name": "chunlin9",
          "author_url": "",
          "post_date": "08/03/2020 08:49:39",
          "content": "<p>Thank you, it looks like by default it is OFF. BTW, where can I find a submit button in the kernel? on the top I can only see \"share\" and \"save version\". Thanks.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 956177,
          "author_name": "luisvivas",
          "author_url": "",
          "post_date": "08/03/2020 10:00:33",
          "content": "<p>Yes, it is off by default. Use the save version button. Then save and run. This will commit and you will be able to see the output, then you can submit.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 956494,
          "author_name": "chunlin9",
          "author_url": "",
          "post_date": "08/03/2020 14:41:50",
          "content": "<p>Thank you!</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "852241": "I'm keep trying to submit a submission file and no matter what I do i'm getting submission errors.\nThe only option that did worked for me was forking the \"global-wheat-detection-first-submission\" kernel. \nI then tried to create a kernel that use the submission file from that kernel and save it as csv file. i double checked with pd.equals that the file is identical and i'm still getting submission error.\n\nI even tried to create a kernel where i manually write the values from that kernel submission file to a new dataframe and save it as cvs, once again I double checked that my new submission.csv file is identical to the original one from the started kernel.\n\nIt's frustrating that there is no explanation of why the submission is failing, any idea how to proceed?",
    "938578": "I had the same issue. First make sure that the submission file is called submission.csv, then make sure that it has the same format as in the sample, it means that it has 2 columns where the first is the file name without jpg and then the string is Accuracy x y width heigth. When there are many objects in the same image, continue the object in the same line, so one image one row.\n\nMy mistake was that I created a file only for submission. Remember that there are hidden images, not only those 10 images that you see in the test folder. So your prediction code needs to be in the same kernel, so the system can generate the code for the rest of the hidden images.\n\nAlso disable internet as it is not allowed to use internet on the notebook that is submitting. If you need a package, crate a dataset or create another notebook and download the packages there, you can use !pip download \"package\" and it will download the wheel files that you need to install. You will need !pip install \"package-name\" \"your-path\"\n\nI hope it helps.",
    "956079": "Hello, you said \"Also disable internet as it is not allowed to use internet on the notebook that is submitting\", I am wondering how possible we could connect to Kaggle if without Internet, we are using laptop, not cell phone.",
    "956086": "No, it does not mean internet in your laptop. It means internet in the Kaggle Notebook. This is done to avoid adding any script downloaded. \n\nThis link explains how to enable internet.\n\nhttps://www.kaggle.com/product-feedback/63544",
    "956115": "Thank you, it looks like by default it is OFF. BTW, where can I find a submit button in the kernel? on the top I can only see \"share\" and \"save version\". Thanks.",
    "956177": "Yes, it is off by default. Use the save version button. Then save and run. This will commit and you will be able to see the output, then you can submit.",
    "956494": "Thank you!"
  },
  "source": "meta"
}