{
  "id": 454121,
  "title": "Submission Actual Test Data",
  "url": "/competitions/child-mind-institute-detect-sleep-states/discussion/454121",
  "author_name": "",
  "post_date": "2023-11-09T03:32:50.087891200Z",
  "votes": 1,
  "comment_count": 5,
  "views": 0,
  "content": "<p>Hi, is it possible to know What is the actual test data size? So I could prepare/anticipate the feature engineering and submission processing. I keep hitting timeout during scoring. </p>",
  "messages": [
    {
      "id": "2518077",
      "postDate": "11/09/2023 03:32:50",
      "content": "<p>Hi, is it possible to know What is the actual test data size? So I could prepare/anticipate the feature engineering and submission processing. I keep hitting timeout during scoring. </p>",
      "rawMarkdown": "Hi, is it possible to know What is the actual test data size? So I could prepare/anticipate the feature engineering and submission processing. I keep hitting timeout during scoring.",
      "votes": null
    },
    {
      "id": "2518101",
      "postDate": "11/09/2023 03:46:28",
      "content": "<p>Yes. In \"Data\" tab <a href=\"https://www.kaggle.com/competitions/child-mind-institute-detect-sleep-states/data\" target=\"_blank\">here</a>, it says</p>\n<blockquote>\n  <p>Note that this is a Code Competition, in which the actual test set is hidden. In this public version, we give some sample data in the correct format to help you author your solutions. The full test set contains about 200 series.</p>\n</blockquote>\n<p>If you wish to test your inference code, try extracting 200 train series and inferring that.</p>",
      "rawMarkdown": "Yes. In \"Data\" tab [here][1], it says\n>Note that this is a Code Competition, in which the actual test set is hidden. In this public version, we give some sample data in the correct format to help you author your solutions. The full test set contains about 200 series.\n\nIf you wish to test your inference code, try extracting 200 train series and inferring that.\n\n[1]: https://www.kaggle.com/competitions/child-mind-institute-detect-sleep-states/data",
      "votes": null
    },
    {
      "id": "2518151",
      "postDate": "11/09/2023 05:07:54",
      "content": "<p>Hi thanks for the reply. The <strong>full</strong> test set is 200 series but the scoring is 200 as well? I thought normally 20% of full test which is around 40 train series.</p>",
      "rawMarkdown": "Hi thanks for the reply. The **full** test set is 200 series but the scoring is 200 as well? I thought normally 20% of full test which is around 40 train series.",
      "votes": null
    },
    {
      "id": "2518192",
      "postDate": "11/09/2023 05:58:51",
      "content": "<p>The public test is 25% and the private is 75%. However our submit notebook <strong>only runs once</strong>. Our submit notebook infers <strong>all 200 series id</strong>. Our submit notebook infers both the public and the private. Afterward, we only see the public LB score. The private LB score is saved by Kaggle. When the competition ends, then Kaggle will immediately show everyone's private LB (because they have all the scores saved already).</p>",
      "rawMarkdown": "The public test is 25% and the private is 75%. However our submit notebook **only runs once**. Our submit notebook infers **all 200 series id**. Our submit notebook infers both the public and the private. Afterward, we only see the public LB score. The private LB score is saved by Kaggle. When the competition ends, then Kaggle will immediately show everyone's private LB (because they have all the scores saved already).",
      "votes": null
    },
    {
      "id": "2518250",
      "postDate": "11/09/2023 06:56:34",
      "content": "<p>I see, thanks for the reply. So more or less 2 minutes per series id</p>",
      "rawMarkdown": "I see, thanks for the reply. So more or less 2 minutes per series id",
      "votes": null
    },
    {
      "id": "2518802",
      "postDate": "11/09/2023 15:01:49",
      "content": "<p>Yes. We have 9 hour limit which is 540 minutes. Therefore we have about 2.7 minutes per series id.</p>",
      "rawMarkdown": "Yes. We have 9 hour limit which is 540 minutes. Therefore we have about 2.7 minutes per series id.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2518101,
      "author_name": "cdeotte",
      "author_url": "",
      "post_date": "11/09/2023 03:46:28",
      "content": "<p>Yes. In \"Data\" tab <a href=\"https://www.kaggle.com/competitions/child-mind-institute-detect-sleep-states/data\" target=\"_blank\">here</a>, it says</p>\n<blockquote>\n  <p>Note that this is a Code Competition, in which the actual test set is hidden. In this public version, we give some sample data in the correct format to help you author your solutions. The full test set contains about 200 series.</p>\n</blockquote>\n<p>If you wish to test your inference code, try extracting 200 train series and inferring that.</p>",
      "votes": null,
      "replies": [
        {
          "id": 2518151,
          "author_name": "rubyhartono",
          "author_url": "",
          "post_date": "11/09/2023 05:07:54",
          "content": "<p>Hi thanks for the reply. The <strong>full</strong> test set is 200 series but the scoring is 200 as well? I thought normally 20% of full test which is around 40 train series.</p>",
          "votes": null,
          "replies": [
            {
              "id": 2518192,
              "author_name": "cdeotte",
              "author_url": "",
              "post_date": "11/09/2023 05:58:51",
              "content": "<p>The public test is 25% and the private is 75%. However our submit notebook <strong>only runs once</strong>. Our submit notebook infers <strong>all 200 series id</strong>. Our submit notebook infers both the public and the private. Afterward, we only see the public LB score. The private LB score is saved by Kaggle. When the competition ends, then Kaggle will immediately show everyone's private LB (because they have all the scores saved already).</p>",
              "votes": null,
              "replies": [
                {
                  "id": 2518250,
                  "author_name": "rubyhartono",
                  "author_url": "",
                  "post_date": "11/09/2023 06:56:34",
                  "content": "<p>I see, thanks for the reply. So more or less 2 minutes per series id</p>",
                  "votes": null,
                  "replies": [
                    {
                      "id": 2518802,
                      "author_name": "cdeotte",
                      "author_url": "",
                      "post_date": "11/09/2023 15:01:49",
                      "content": "<p>Yes. We have 9 hour limit which is 540 minutes. Therefore we have about 2.7 minutes per series id.</p>",
                      "votes": null,
                      "replies": []
                    }
                  ]
                }
              ]
            }
          ]
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2518077": "Hi, is it possible to know What is the actual test data size? So I could prepare/anticipate the feature engineering and submission processing. I keep hitting timeout during scoring.",
    "2518101": "Yes. In \"Data\" tab [here][1], it says\n>Note that this is a Code Competition, in which the actual test set is hidden. In this public version, we give some sample data in the correct format to help you author your solutions. The full test set contains about 200 series.\n\nIf you wish to test your inference code, try extracting 200 train series and inferring that.\n\n[1]: https://www.kaggle.com/competitions/child-mind-institute-detect-sleep-states/data",
    "2518151": "Hi thanks for the reply. The **full** test set is 200 series but the scoring is 200 as well? I thought normally 20% of full test which is around 40 train series.",
    "2518192": "The public test is 25% and the private is 75%. However our submit notebook **only runs once**. Our submit notebook infers **all 200 series id**. Our submit notebook infers both the public and the private. Afterward, we only see the public LB score. The private LB score is saved by Kaggle. When the competition ends, then Kaggle will immediately show everyone's private LB (because they have all the scores saved already).",
    "2518250": "I see, thanks for the reply. So more or less 2 minutes per series id",
    "2518802": "Yes. We have 9 hour limit which is 540 minutes. Therefore we have about 2.7 minutes per series id."
  },
  "source": "meta"
}