{
  "id": 72423,
  "title": "Kaggle Kernel no Output",
  "url": "/competitions/quora-insincere-questions-classification/discussion/72423",
  "author_name": "",
  "post_date": "2018-11-23T05:47:07.548006600Z",
  "votes": null,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Hi All,</p>\n\n<p>I have a kernel where I do my prediction.\nI use my own pretrained models that I trained outside of the kernel (training takes too long for the Kernel). I than use that to predict on test data and output  submission file.</p>\n\n<p>However, when I write the predictions to csv there is no outputs tab in my kernel. I am stuck on why this is. Is it because I am not allowed to use pretrained models files I uploaded or am I missing something?</p>",
  "messages": [
    {
      "id": "426350",
      "postDate": "11/23/2018 05:47:07",
      "content": "<p>Hi All,</p>\n\n<p>I have a kernel where I do my prediction.\nI use my own pretrained models that I trained outside of the kernel (training takes too long for the Kernel). I than use that to predict on test data and output  submission file.</p>\n\n<p>However, when I write the predictions to csv there is no outputs tab in my kernel. I am stuck on why this is. Is it because I am not allowed to use pretrained models files I uploaded or am I missing something?</p>",
      "rawMarkdown": "Hi All,\n\nI have a kernel where I do my prediction.\nI use my own pretrained models that I trained outside of the kernel (training takes too long for the Kernel). I than use that to predict on test data and output  submission file.\n\nHowever, when I write the predictions to csv there is no outputs tab in my kernel. I am stuck on why this is. Is it because I am not allowed to use pretrained models files I uploaded or am I missing something?",
      "votes": null
    },
    {
      "id": "426362",
      "postDate": "11/23/2018 06:25:00",
      "content": "<p>I guess it's the problem of your code. You can use additional data outside but the submit button will not appear, instead of not having an output file.</p>",
      "rawMarkdown": "I guess it's the problem of your code. You can use additional data outside but the submit button will not appear, instead of not having an output file.",
      "votes": null
    },
    {
      "id": "426448",
      "postDate": "11/23/2018 09:23:33",
      "content": "<p>I couldn't submit prediction file (although it was properly prepared and named submission.csv) when I also saved some model parameters in another csv file. It seems that beside the condition on the output file there is another one that it must be the only output of the kernel.\nBut I also remember in one other competition the submitting was disabled automatically if there was external data used in my kernel. Could it be this case?</p>",
      "rawMarkdown": "I couldn't submit prediction file (although it was properly prepared and named submission.csv) when I also saved some model parameters in another csv file. It seems that beside the condition on the output file there is another one that it must be the only output of the kernel.\nBut I also remember in one other competition the submitting was disabled automatically if there was external data used in my kernel. Could it be this case?",
      "votes": null
    },
    {
      "id": "427308",
      "postDate": "11/25/2018 06:51:32",
      "content": "<p>I tried a super naive kernel where I just upload my predictions (local) and load into pandas dataframe and output. This works, however, unable to submit as it says I use an external data source blah2. </p>\n\n<p>Seems like I will have to create and chain up multiple kernels. I wish this was easier that how it is done ......</p>",
      "rawMarkdown": "I tried a super naive kernel where I just upload my predictions (local) and load into pandas dataframe and output. This works, however, unable to submit as it says I use an external data source blah2. \n\nSeems like I will have to create and chain up multiple kernels. I wish this was easier that how it is done ......",
      "votes": null
    },
    {
      "id": "427836",
      "postDate": "11/26/2018 08:40:32",
      "content": "<p>Isn't chaining kernels counted as a method of using external data?</p>",
      "rawMarkdown": "Isn't chaining kernels counted as a method of using external data?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 426362,
      "author_name": "melmew",
      "author_url": "",
      "post_date": "11/23/2018 06:25:00",
      "content": "<p>I guess it's the problem of your code. You can use additional data outside but the submit button will not appear, instead of not having an output file.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 426448,
      "author_name": "akuropatwinski",
      "author_url": "",
      "post_date": "11/23/2018 09:23:33",
      "content": "<p>I couldn't submit prediction file (although it was properly prepared and named submission.csv) when I also saved some model parameters in another csv file. It seems that beside the condition on the output file there is another one that it must be the only output of the kernel.\nBut I also remember in one other competition the submitting was disabled automatically if there was external data used in my kernel. Could it be this case?</p>",
      "votes": null,
      "replies": [
        {
          "id": 427308,
          "author_name": "mhs123",
          "author_url": "",
          "post_date": "11/25/2018 06:51:32",
          "content": "<p>I tried a super naive kernel where I just upload my predictions (local) and load into pandas dataframe and output. This works, however, unable to submit as it says I use an external data source blah2. </p>\n\n<p>Seems like I will have to create and chain up multiple kernels. I wish this was easier that how it is done ......</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 427836,
          "author_name": "akuropatwinski",
          "author_url": "",
          "post_date": "11/26/2018 08:40:32",
          "content": "<p>Isn't chaining kernels counted as a method of using external data?</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "426350": "Hi All,\n\nI have a kernel where I do my prediction.\nI use my own pretrained models that I trained outside of the kernel (training takes too long for the Kernel). I than use that to predict on test data and output  submission file.\n\nHowever, when I write the predictions to csv there is no outputs tab in my kernel. I am stuck on why this is. Is it because I am not allowed to use pretrained models files I uploaded or am I missing something?",
    "426362": "I guess it's the problem of your code. You can use additional data outside but the submit button will not appear, instead of not having an output file.",
    "426448": "I couldn't submit prediction file (although it was properly prepared and named submission.csv) when I also saved some model parameters in another csv file. It seems that beside the condition on the output file there is another one that it must be the only output of the kernel.\nBut I also remember in one other competition the submitting was disabled automatically if there was external data used in my kernel. Could it be this case?",
    "427308": "I tried a super naive kernel where I just upload my predictions (local) and load into pandas dataframe and output. This works, however, unable to submit as it says I use an external data source blah2. \n\nSeems like I will have to create and chain up multiple kernels. I wish this was easier that how it is done ......",
    "427836": "Isn't chaining kernels counted as a method of using external data?"
  },
  "source": "meta"
}