{
  "id": 125186,
  "title": "Notebook Timeout Issue",
  "url": "/competitions/bengaliai-cv19/discussion/125186",
  "author_name": "",
  "post_date": "2020-01-09T05:05:27.837030600Z",
  "votes": 1,
  "comment_count": 4,
  "views": 0,
  "content": "<p>My kernel is running successfully in less than 5 minutes when commiting. But when submitting to the competition it's running for 2 hours and showing 'Notebook Timeout' error. What may be the possible reason of this?</p>",
  "messages": [
    {
      "id": "714129",
      "postDate": "01/09/2020 05:05:27",
      "content": "<p>My kernel is running successfully in less than 5 minutes when commiting. But when submitting to the competition it's running for 2 hours and showing 'Notebook Timeout' error. What may be the possible reason of this?</p>",
      "rawMarkdown": "My kernel is running successfully in less than 5 minutes when commiting. But when submitting to the competition it's running for 2 hours and showing 'Notebook Timeout' error. What may be the possible reason of this?",
      "votes": null
    },
    {
      "id": "714164",
      "postDate": "01/09/2020 06:12:14",
      "content": "<p>Because the test dataset you see is not the real data . The test data is quite large . To process it faster you need to have right number of batch_size and right way of processing the test loader else it will take long time . You can predict the performance by taking 10000 rows from the train loader and try to predict and commit using your model .  </p>",
      "rawMarkdown": "Because the test dataset you see is not the real data . The test data is quite large . To process it faster you need to have right number of batch_size and right way of processing the test loader else it will take long time . You can predict the performance by taking 10000 rows from the train loader and try to predict and commit using your model .",
      "votes": null
    },
    {
      "id": "714291",
      "postDate": "01/09/2020 09:20:15",
      "content": "<p>Set Batch_size = 128</p>",
      "rawMarkdown": "Set Batch_size = 128",
      "votes": null
    },
    {
      "id": "714354",
      "postDate": "01/09/2020 10:40:15",
      "content": "<p>Thanks a lot. I've figured out the issue and it's working.</p>",
      "rawMarkdown": "Thanks a lot. I've figured out the issue and it's working.",
      "votes": null
    },
    {
      "id": "749157",
      "postDate": "02/18/2020 11:28:01",
      "content": "<p>How do you solve it?</p>",
      "rawMarkdown": "How do you solve it?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 714164,
      "author_name": "phoenix9032",
      "author_url": "",
      "post_date": "01/09/2020 06:12:14",
      "content": "<p>Because the test dataset you see is not the real data . The test data is quite large . To process it faster you need to have right number of batch_size and right way of processing the test loader else it will take long time . You can predict the performance by taking 10000 rows from the train loader and try to predict and commit using your model .  </p>",
      "votes": null,
      "replies": [
        {
          "id": 714354,
          "author_name": "adnantoky",
          "author_url": "",
          "post_date": "01/09/2020 10:40:15",
          "content": "<p>Thanks a lot. I've figured out the issue and it's working.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 749157,
          "author_name": "supererstar",
          "author_url": "",
          "post_date": "02/18/2020 11:28:01",
          "content": "<p>How do you solve it?</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 714291,
      "author_name": "karan07",
      "author_url": "",
      "post_date": "01/09/2020 09:20:15",
      "content": "<p>Set Batch_size = 128</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "714129": "My kernel is running successfully in less than 5 minutes when commiting. But when submitting to the competition it's running for 2 hours and showing 'Notebook Timeout' error. What may be the possible reason of this?",
    "714164": "Because the test dataset you see is not the real data . The test data is quite large . To process it faster you need to have right number of batch_size and right way of processing the test loader else it will take long time . You can predict the performance by taking 10000 rows from the train loader and try to predict and commit using your model .",
    "714291": "Set Batch_size = 128",
    "714354": "Thanks a lot. I've figured out the issue and it's working.",
    "749157": "How do you solve it?"
  },
  "source": "meta"
}