{
  "id": 160204,
  "title": "Questions about Test Set & Pre-trained Model",
  "url": "/competitions/birdsong-recognition/discussion/160204",
  "author_name": "",
  "post_date": "2020-06-20T08:34:52.753398600Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I have 2 questions\n1. What does the author mean with this line: <code>site 3 has relatively few rows in the test set and needs lower time resolution predictions</code>?\n2. Can I use pre-trained model?</p>\n\n<p>Sorry if it sounds noob. But it's my first time to join any active competition. Thank you so much for helping.</p>",
  "messages": [
    {
      "id": "894218",
      "postDate": "06/20/2020 08:34:52",
      "content": "<p>I have 2 questions\n1. What does the author mean with this line: <code>site 3 has relatively few rows in the test set and needs lower time resolution predictions</code>?\n2. Can I use pre-trained model?</p>\n\n<p>Sorry if it sounds noob. But it's my first time to join any active competition. Thank you so much for helping.</p>",
      "rawMarkdown": "I have 2 questions\n1. What does the author mean with this line: `site 3 has relatively few rows in the test set and needs lower time resolution predictions`?\n2. Can I use pre-trained model?\n\nSorry if it sounds noob. But it's my first time to join any active competition. Thank you so much for helping.",
      "votes": null
    },
    {
      "id": "894873",
      "postDate": "06/20/2020 21:00:48",
      "content": "<ol>\n<li><p>All test audios are 10 mins, irrespective of the site. For site 1 and site 2, labels are given for every 5 second window, which means there are (60*10)/5 = 120 rows for each audio file (of site 1 &amp; 2). For site 3, labels are only given at file level, which means that for every 10 min audio file of site 3, there is 1 corresponding row with bird labels.</p></li>\n<li><p>Yes, pre-trained models are allowed.</p></li>\n</ol>",
      "rawMarkdown": "1. All test audios are 10 mins, irrespective of the site. For site 1 and site 2, labels are given for every 5 second window, which means there are (60*10)/5 = 120 rows for each audio file (of site 1 &amp; 2). For site 3, labels are only given at file level, which means that for every 10 min audio file of site 3, there is 1 corresponding row with bird labels.\n\n2. Yes, pre-trained models are allowed.",
      "votes": null
    },
    {
      "id": "896404",
      "postDate": "06/22/2020 06:40:04",
      "content": "<p>Thanks for the answer</p>",
      "rawMarkdown": "Thanks for the answer",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 894873,
      "author_name": "dhruvrnaik",
      "author_url": "",
      "post_date": "06/20/2020 21:00:48",
      "content": "<ol>\n<li><p>All test audios are 10 mins, irrespective of the site. For site 1 and site 2, labels are given for every 5 second window, which means there are (60*10)/5 = 120 rows for each audio file (of site 1 &amp; 2). For site 3, labels are only given at file level, which means that for every 10 min audio file of site 3, there is 1 corresponding row with bird labels.</p></li>\n<li><p>Yes, pre-trained models are allowed.</p></li>\n</ol>",
      "votes": null,
      "replies": [
        {
          "id": 896404,
          "author_name": "dipta007",
          "author_url": "",
          "post_date": "06/22/2020 06:40:04",
          "content": "<p>Thanks for the answer</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "894218": "I have 2 questions\n1. What does the author mean with this line: `site 3 has relatively few rows in the test set and needs lower time resolution predictions`?\n2. Can I use pre-trained model?\n\nSorry if it sounds noob. But it's my first time to join any active competition. Thank you so much for helping.",
    "894873": "1. All test audios are 10 mins, irrespective of the site. For site 1 and site 2, labels are given for every 5 second window, which means there are (60*10)/5 = 120 rows for each audio file (of site 1 &amp; 2). For site 3, labels are only given at file level, which means that for every 10 min audio file of site 3, there is 1 corresponding row with bird labels.\n\n2. Yes, pre-trained models are allowed.",
    "896404": "Thanks for the answer"
  },
  "source": "meta"
}