{
  "id": 208386,
  "title": "Submission confusion",
  "url": "/competitions/cassava-leaf-disease-classification/discussion/208386",
  "author_name": "",
  "post_date": "2021-01-03T07:59:45.558074200Z",
  "votes": null,
  "comment_count": 2,
  "views": 0,
  "content": "<p>There seems to be only one test image for public leaderboard. In that cases accuracy should be 0 or 100% but on leaderboard it is 0.8-0.9. How can this be? How are we supposed to benchmark with just one test image?</p>",
  "messages": [
    {
      "id": "1136573",
      "postDate": "01/03/2021 07:59:45",
      "content": "<p>There seems to be only one test image for public leaderboard. In that cases accuracy should be 0 or 100% but on leaderboard it is 0.8-0.9. How can this be? How are we supposed to benchmark with just one test image?</p>",
      "rawMarkdown": "There seems to be only one test image for public leaderboard. In that cases accuracy should be 0 or 100% but on leaderboard it is 0.8-0.9. How can this be? How are we supposed to benchmark with just one test image?",
      "votes": null
    },
    {
      "id": "1136638",
      "postDate": "01/03/2021 09:33:32",
      "content": "<p>Hi,<br>\nThe image given in the test is just to complete the pipeline. After submission, the code will be run against a bigger set of unseen images (~15K) and the resulting score will be the one shown on the LB.</p>",
      "rawMarkdown": "Hi,\nThe image given in the test is just to complete the pipeline. After submission, the code will be run against a bigger set of unseen images (~15K) and the resulting score will be the one shown on the LB.",
      "votes": null
    },
    {
      "id": "1136769",
      "postDate": "01/03/2021 12:07:48",
      "content": "<p>From the leaderboard tab on the competition page</p>\n<blockquote>\n  <p>This leaderboard is calculated with approximately 31% of the test data.</p>\n  <p>The final results will be based on the other 69%, so the final standings may be different.</p>\n</blockquote>",
      "rawMarkdown": "From the leaderboard tab on the competition page\n\n> This leaderboard is calculated with approximately 31% of the test data.\n\n> The final results will be based on the other 69%, so the final standings may be different.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1136638,
      "author_name": "ashwanikumar04",
      "author_url": "",
      "post_date": "01/03/2021 09:33:32",
      "content": "<p>Hi,<br>\nThe image given in the test is just to complete the pipeline. After submission, the code will be run against a bigger set of unseen images (~15K) and the resulting score will be the one shown on the LB.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1136769,
      "author_name": "mohitve",
      "author_url": "",
      "post_date": "01/03/2021 12:07:48",
      "content": "<p>From the leaderboard tab on the competition page</p>\n<blockquote>\n  <p>This leaderboard is calculated with approximately 31% of the test data.</p>\n  <p>The final results will be based on the other 69%, so the final standings may be different.</p>\n</blockquote>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1136573": "There seems to be only one test image for public leaderboard. In that cases accuracy should be 0 or 100% but on leaderboard it is 0.8-0.9. How can this be? How are we supposed to benchmark with just one test image?",
    "1136638": "Hi,\nThe image given in the test is just to complete the pipeline. After submission, the code will be run against a bigger set of unseen images (~15K) and the resulting score will be the one shown on the LB.",
    "1136769": "From the leaderboard tab on the competition page\n\n> This leaderboard is calculated with approximately 31% of the test data.\n\n> The final results will be based on the other 69%, so the final standings may be different."
  },
  "source": "meta"
}