{
  "id": 486100,
  "title": "Test submission inquiry ",
  "url": "/competitions/hms-harmful-brain-activity-classification/discussion/486100",
  "author_name": "",
  "post_date": "2024-03-23T15:02:04.682269800Z",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Just want to confirm if we are building the model to only test for one row of eeg_id for the final submission as it seems that’s the case with the test data that’s there. Unless I am missing something, should the submission.csv that we provide only have one row of data? Also, does it matter how much decimal the values are for the vote probability?</p>",
  "messages": [
    {
      "id": "2712448",
      "postDate": "03/23/2024 15:02:04",
      "content": "<p>Just want to confirm if we are building the model to only test for one row of eeg_id for the final submission as it seems that’s the case with the test data that’s there. Unless I am missing something, should the submission.csv that we provide only have one row of data? Also, does it matter how much decimal the values are for the vote probability?</p>",
      "rawMarkdown": "Just want to confirm if we are building the model to only test for one row of eeg_id for the final submission as it seems that’s the case with the test data that’s there. Unless I am missing something, should the submission.csv that we provide only have one row of data? Also, does it matter how much decimal the values are for the vote probability?",
      "votes": null
    },
    {
      "id": "2712598",
      "postDate": "03/23/2024 17:41:19",
      "content": "<p>The test data that we can download and see is not the real test data. When we submit our code, the test data gets replaced with the real test data (which is about 2500 rows) and our code \"sees\" the real test data and makes ~2500 predictions.</p>",
      "rawMarkdown": "The test data that we can download and see is not the real test data. When we submit our code, the test data gets replaced with the real test data (which is about 2500 rows) and our code \"sees\" the real test data and makes ~2500 predictions.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2712598,
      "author_name": "cdeotte",
      "author_url": "",
      "post_date": "03/23/2024 17:41:19",
      "content": "<p>The test data that we can download and see is not the real test data. When we submit our code, the test data gets replaced with the real test data (which is about 2500 rows) and our code \"sees\" the real test data and makes ~2500 predictions.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2712448": "Just want to confirm if we are building the model to only test for one row of eeg_id for the final submission as it seems that’s the case with the test data that’s there. Unless I am missing something, should the submission.csv that we provide only have one row of data? Also, does it matter how much decimal the values are for the vote probability?",
    "2712598": "The test data that we can download and see is not the real test data. When we submit our code, the test data gets replaced with the real test data (which is about 2500 rows) and our code \"sees\" the real test data and makes ~2500 predictions."
  },
  "source": "meta"
}