{
  "id": 457122,
  "title": "Test data/ Loader/ Path",
  "url": "/competitions/UBC-OCEAN/discussion/457122",
  "author_name": "AhmedBil",
  "post_date": "2023-11-23T05:56:38.364000",
  "votes": 3,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Hi all,<br>\nI am curious about how they are going to evaluate the notebook/code for 22% of the data. As there is no global variable (for actual test data), no path to run against DL model. Any suggestions/points I am missing here. Thanks.</p>",
  "messages": [
    {
      "id": 2535696,
      "postDate": "2023-11-23T14:38:34.543Z",
      "content": "<p>My understanding on how it works is that when you run the notebook on the test data (1 record)<br>\nWhen it gets submitted it changes the test data file to the one containing ~2000 records.<br>\nThe public leaderboard score you see corresponds to 22% of those 2000 records.<br>\nSo in submission the test.csv is different than the one you see when writing you're notebook</p>",
      "rawMarkdown": "My understanding on how it works is that when you run the notebook on the test data (1 record)\nWhen it gets submitted it changes the test data file to the one containing ~2000 records.\nThe public leaderboard score you see corresponds to 22% of those 2000 records.\nSo in submission the test.csv is different than the one you see when writing you're notebook",
      "votes": 3,
      "replies": [
        {
          "id": 2536267,
          "postDate": "2023-11-24T04:45:56.423Z",
          "content": "<p>Thanks. That was really informative. Would you be able to visualize the test dataset (images) post-submission?</p>",
          "rawMarkdown": "Thanks. That was really informative. Would you be able to visualize the test dataset (images) post-submission?",
          "replies": [
            {
              "id": 2536729,
              "postDate": "2023-11-24T12:10:17.780Z",
              "content": "<p>No, since you don't get to view the submission notebook.<br>\nAll kaggle will tell you is what error the notebook threw, or the score</p>",
              "rawMarkdown": "No, since you don't get to view the submission notebook.\nAll kaggle will tell you is what error the notebook threw, or the score",
              "votes": 1
            },
            {
              "id": 2540992,
              "postDate": "2023-11-28T06:19:51.573Z",
              "content": "<p>Thanks for your inputs. Appreciate it.</p>",
              "rawMarkdown": "Thanks for your inputs. Appreciate it."
            }
          ]
        }
      ]
    },
    {
      "id": 2535145,
      "postDate": "2023-11-23T05:56:38.363Z",
      "content": "<p>Hi all,<br>\nI am curious about how they are going to evaluate the notebook/code for 22% of the data. As there is no global variable (for actual test data), no path to run against DL model. Any suggestions/points I am missing here. Thanks.</p>",
      "rawMarkdown": "Hi all,\nI am curious about how they are going to evaluate the notebook/code for 22% of the data. As there is no global variable (for actual test data), no path to run against DL model. Any suggestions/points I am missing here. Thanks.",
      "votes": 3
    }
  ],
  "comments": [
    {
      "id": 2535696,
      "author_name": "KSMCG90",
      "author_url": "",
      "post_date": "2023-11-23T14:38:34.543000",
      "content": "<p>My understanding on how it works is that when you run the notebook on the test data (1 record)<br>\nWhen it gets submitted it changes the test data file to the one containing ~2000 records.<br>\nThe public leaderboard score you see corresponds to 22% of those 2000 records.<br>\nSo in submission the test.csv is different than the one you see when writing you're notebook</p>",
      "votes": 3,
      "replies": [
        {
          "id": 2536267,
          "author_name": "AhmedBil",
          "author_url": "",
          "post_date": "2023-11-24T04:45:56.423000",
          "content": "<p>Thanks. That was really informative. Would you be able to visualize the test dataset (images) post-submission?</p>",
          "votes": 0,
          "replies": [
            {
              "id": 2536729,
              "author_name": "KSMCG90",
              "author_url": "",
              "post_date": "2023-11-24T12:10:17.780000",
              "content": "<p>No, since you don't get to view the submission notebook.<br>\nAll kaggle will tell you is what error the notebook threw, or the score</p>",
              "votes": 1,
              "replies": []
            },
            {
              "id": 2540992,
              "author_name": "AhmedBil",
              "author_url": "",
              "post_date": "2023-11-28T06:19:51.573000",
              "content": "<p>Thanks for your inputs. Appreciate it.</p>",
              "votes": 0,
              "replies": []
            }
          ]
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2535696": "My understanding on how it works is that when you run the notebook on the test data (1 record)\nWhen it gets submitted it changes the test data file to the one containing ~2000 records.\nThe public leaderboard score you see corresponds to 22% of those 2000 records.\nSo in submission the test.csv is different than the one you see when writing you're notebook",
    "2535145": "Hi all,\nI am curious about how they are going to evaluate the notebook/code for 22% of the data. As there is no global variable (for actual test data), no path to run against DL model. Any suggestions/points I am missing here. Thanks."
  }
}