{
  "id": 159121,
  "title": "I like the sound of birds but what is up with this test file?",
  "url": "/competitions/birdsong-recognition/discussion/159121",
  "author_name": "",
  "post_date": "2020-06-16T14:02:36.766975700Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>The competition is self-understood. But, am I the only one who is finding it troublesome to generate my first submission file. </p>\n\n<p>The test set has only three rows and the rest are said to be hidden.</p>\n\n<p>Does this mean my submission notebook needs to predict only the three birds in the given test set perfectly to get a high LB?</p>",
  "messages": [
    {
      "id": "888687",
      "postDate": "06/16/2020 14:02:36",
      "content": "<p>The competition is self-understood. But, am I the only one who is finding it troublesome to generate my first submission file. </p>\n\n<p>The test set has only three rows and the rest are said to be hidden.</p>\n\n<p>Does this mean my submission notebook needs to predict only the three birds in the given test set perfectly to get a high LB?</p>",
      "rawMarkdown": "The competition is self-understood. But, am I the only one who is finding it troublesome to generate my first submission file. \n\nThe test set has only three rows and the rest are said to be hidden.\n\nDoes this mean my submission notebook needs to predict only the three birds in the given test set perfectly to get a high LB?",
      "votes": null
    },
    {
      "id": "888859",
      "postDate": "06/16/2020 15:47:18",
      "content": "<p>The 3 rows are given as an example so that we can write our inference pipeline accordingly.\nSince the <code>test_audio</code> file is hidden, you can't predictions for the 3 <code>row_id</code>s in the <code>test.csv</code> file. </p>\n\n<p>I am guessing the predictions on test_audio will only happen during kernel re-run on submission.</p>",
      "rawMarkdown": "The 3 rows are given as an example so that we can write our inference pipeline accordingly.\nSince the `test_audio` file is hidden, you can't predictions for the 3 `row_id`s in the `test.csv` file. \n\nI am guessing the predictions on test_audio will only happen during kernel re-run on submission.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 888859,
      "author_name": "dhruvrnaik",
      "author_url": "",
      "post_date": "06/16/2020 15:47:18",
      "content": "<p>The 3 rows are given as an example so that we can write our inference pipeline accordingly.\nSince the <code>test_audio</code> file is hidden, you can't predictions for the 3 <code>row_id</code>s in the <code>test.csv</code> file. </p>\n\n<p>I am guessing the predictions on test_audio will only happen during kernel re-run on submission.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "888687": "The competition is self-understood. But, am I the only one who is finding it troublesome to generate my first submission file. \n\nThe test set has only three rows and the rest are said to be hidden.\n\nDoes this mean my submission notebook needs to predict only the three birds in the given test set perfectly to get a high LB?",
    "888859": "The 3 rows are given as an example so that we can write our inference pipeline accordingly.\nSince the `test_audio` file is hidden, you can't predictions for the 3 `row_id`s in the `test.csv` file. \n\nI am guessing the predictions on test_audio will only happen during kernel re-run on submission."
  },
  "source": "meta"
}