{
  "id": 48256,
  "title": "Please Help: How to Overcome Kaggle Kernel 1 hour Runtime Constraint",
  "url": "/competitions/sp-society-camera-model-identification/discussion/48256",
  "author_name": "",
  "post_date": "2018-01-25T09:20:59.705542400Z",
  "votes": null,
  "comment_count": 3,
  "views": 0,
  "content": "<p>While running my script in Kaggle kernel, after one hour the process was terminated. I discovered that Kaggle kernels has a constraint that the maximum runtime is 1 hour. \nI don't know how to overcome this situation? My script takes more than one hour. </p>",
  "messages": [
    {
      "id": "273839",
      "postDate": "01/25/2018 09:20:59",
      "content": "<p>While running my script in Kaggle kernel, after one hour the process was terminated. I discovered that Kaggle kernels has a constraint that the maximum runtime is 1 hour. \nI don't know how to overcome this situation? My script takes more than one hour. </p>",
      "rawMarkdown": "While running my script in Kaggle kernel, after one hour the process was terminated. I discovered that Kaggle kernels has a constraint that the maximum runtime is 1 hour. \nI don't know how to overcome this situation? My script takes more than one hour.",
      "votes": null
    },
    {
      "id": "273841",
      "postDate": "01/25/2018 09:35:27",
      "content": "<p>Hi  </p>\n\n<p>You can divide your kernel , such as using your first kernel create an output file ,and  then use that file in your second kernel as an input.</p>\n\n<p>Example:\n<a href=\"https://www.kaggle.com/szamil/where-is-my-output-file\">https://www.kaggle.com/szamil/where-is-my-output-file</a></p>",
      "rawMarkdown": "Hi  \n\nYou can divide your kernel , such as using your first kernel create an output file ,and  then use that file in your second kernel as an input.\n\nExample:\nhttps://www.kaggle.com/szamil/where-is-my-output-file",
      "votes": null
    },
    {
      "id": "273932",
      "postDate": "01/25/2018 14:24:23",
      "content": "<p>Thanks a lot :)</p>",
      "rawMarkdown": "Thanks a lot :)",
      "votes": null
    },
    {
      "id": "273936",
      "postDate": "01/25/2018 14:29:38",
      "content": "<p>For the sake of documentation I will write the full steps I followed to solve the problem. \nYou need to create multiple kernels and the results of each kernel will be the input of the next kernel. </p>\n\n<p>Kaggle after each run cleans the output file so you can't run the same kernel code twice with commenting the first part and run the second part because the results of the first  part will be cleaned up if it is in the output folder.  </p>\n\n<p>Also if you used AddDataSource option and pointed to the file produced from the first run you will survive and do the commenting trick I mentioned above but when you submit your results, all the added data source are removed I don't know why? </p>\n\n<p>So in conclusion the best work around I reached till now, is to make multiple kernels and in the next kernel you add data source from the past kernel, even when you submit your results and the data source is cleaned, you can easily refer again to the past kernel as your data source and in include it in your current kernel. </p>",
      "rawMarkdown": "For the sake of documentation I will write the full steps I followed to solve the problem. \nYou need to create multiple kernels and the results of each kernel will be the input of the next kernel. \n\nKaggle after each run cleans the output file so you can't run the same kernel code twice with commenting the first part and run the second part because the results of the first  part will be cleaned up if it is in the output folder.  \n\nAlso if you used AddDataSource option and pointed to the file produced from the first run you will survive and do the commenting trick I mentioned above but when you submit your results, all the added data source are removed I don't know why? \n\nSo in conclusion the best work around I reached till now, is to make multiple kernels and in the next kernel you add data source from the past kernel, even when you submit your results and the data source is cleaned, you can easily refer again to the past kernel as your data source and in include it in your current kernel.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 273841,
      "author_name": "ozgurb",
      "author_url": "",
      "post_date": "01/25/2018 09:35:27",
      "content": "<p>Hi  </p>\n\n<p>You can divide your kernel , such as using your first kernel create an output file ,and  then use that file in your second kernel as an input.</p>\n\n<p>Example:\n<a href=\"https://www.kaggle.com/szamil/where-is-my-output-file\">https://www.kaggle.com/szamil/where-is-my-output-file</a></p>",
      "votes": null,
      "replies": [
        {
          "id": 273932,
          "author_name": "thegenuismentalist",
          "author_url": "",
          "post_date": "01/25/2018 14:24:23",
          "content": "<p>Thanks a lot :)</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 273936,
      "author_name": "thegenuismentalist",
      "author_url": "",
      "post_date": "01/25/2018 14:29:38",
      "content": "<p>For the sake of documentation I will write the full steps I followed to solve the problem. \nYou need to create multiple kernels and the results of each kernel will be the input of the next kernel. </p>\n\n<p>Kaggle after each run cleans the output file so you can't run the same kernel code twice with commenting the first part and run the second part because the results of the first  part will be cleaned up if it is in the output folder.  </p>\n\n<p>Also if you used AddDataSource option and pointed to the file produced from the first run you will survive and do the commenting trick I mentioned above but when you submit your results, all the added data source are removed I don't know why? </p>\n\n<p>So in conclusion the best work around I reached till now, is to make multiple kernels and in the next kernel you add data source from the past kernel, even when you submit your results and the data source is cleaned, you can easily refer again to the past kernel as your data source and in include it in your current kernel. </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "273839": "While running my script in Kaggle kernel, after one hour the process was terminated. I discovered that Kaggle kernels has a constraint that the maximum runtime is 1 hour. \nI don't know how to overcome this situation? My script takes more than one hour.",
    "273841": "Hi  \n\nYou can divide your kernel , such as using your first kernel create an output file ,and  then use that file in your second kernel as an input.\n\nExample:\nhttps://www.kaggle.com/szamil/where-is-my-output-file",
    "273932": "Thanks a lot :)",
    "273936": "For the sake of documentation I will write the full steps I followed to solve the problem. \nYou need to create multiple kernels and the results of each kernel will be the input of the next kernel. \n\nKaggle after each run cleans the output file so you can't run the same kernel code twice with commenting the first part and run the second part because the results of the first  part will be cleaned up if it is in the output folder.  \n\nAlso if you used AddDataSource option and pointed to the file produced from the first run you will survive and do the commenting trick I mentioned above but when you submit your results, all the added data source are removed I don't know why? \n\nSo in conclusion the best work around I reached till now, is to make multiple kernels and in the next kernel you add data source from the past kernel, even when you submit your results and the data source is cleaned, you can easily refer again to the past kernel as your data source and in include it in your current kernel."
  },
  "source": "meta"
}