{
  "id": 238097,
  "title": "Update - Global Wheat Challenge 2021 - New version of the dataset !",
  "url": "/competitions/global-wheat-detection/discussion/238097",
  "author_name": "Etienne David",
  "post_date": "2021-05-11T07:53:36.942000",
  "votes": 2,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Hello to everybody !</p>\n<p>Thanks a lot for your work and feedbacks since last years. We are happy to share with you the second version of the dataset + an update of the benchmark rules:</p>\n<ul>\n<li>We have added more than 120k coming from even much more diverse conditions: this help to get a better test set </li>\n<li>Labels have been corrected with an annotation company</li>\n<li>We have change the metric: matching is done for IoU = 0.5 only but now it's averaged per domain</li>\n<li>All models (including Yolov5) are allowed !</li>\n</ul>\n<p>More informations on the website for the coming challenge : <a href=\"http://www.global-wheat.com/\" target=\"_blank\">http://www.global-wheat.com/</a> </p>\n<p>We hope that you will be able to transfer your innovation there :) !</p>",
  "messages": [
    {
      "id": 1301711,
      "postDate": "2021-05-11T07:53:36.943Z",
      "content": "<p>Hello to everybody !</p>\n<p>Thanks a lot for your work and feedbacks since last years. We are happy to share with you the second version of the dataset + an update of the benchmark rules:</p>\n<ul>\n<li>We have added more than 120k coming from even much more diverse conditions: this help to get a better test set </li>\n<li>Labels have been corrected with an annotation company</li>\n<li>We have change the metric: matching is done for IoU = 0.5 only but now it's averaged per domain</li>\n<li>All models (including Yolov5) are allowed !</li>\n</ul>\n<p>More informations on the website for the coming challenge : <a href=\"http://www.global-wheat.com/\" target=\"_blank\">http://www.global-wheat.com/</a> </p>\n<p>We hope that you will be able to transfer your innovation there :) !</p>",
      "rawMarkdown": "Hello to everybody !\n\nThanks a lot for your work and feedbacks since last years. We are happy to share with you the second version of the dataset + an update of the benchmark rules:\n- We have added more than 120k coming from even much more diverse conditions: this help to get a better test set \n- Labels have been corrected with an annotation company\n- We have change the metric: matching is done for IoU = 0.5 only but now it's averaged per domain\n- All models (including Yolov5) are allowed !\n\nMore informations on the website for the coming challenge : http://www.global-wheat.com/ \n\nWe hope that you will be able to transfer your innovation there :) !",
      "votes": 1
    },
    {
      "id": 1302232,
      "postDate": "2021-05-11T12:55:09.733Z",
      "rawMarkdown": "",
      "isDeleted": true
    }
  ],
  "comments": [
    {
      "id": 1302232,
      "author_name": "",
      "author_url": "",
      "post_date": "2021-05-11T12:55:09.733000",
      "content": "",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1301711": "Hello to everybody !\n\nThanks a lot for your work and feedbacks since last years. We are happy to share with you the second version of the dataset + an update of the benchmark rules:\n- We have added more than 120k coming from even much more diverse conditions: this help to get a better test set \n- Labels have been corrected with an annotation company\n- We have change the metric: matching is done for IoU = 0.5 only but now it's averaged per domain\n- All models (including Yolov5) are allowed !\n\nMore informations on the website for the coming challenge : http://www.global-wheat.com/ \n\nWe hope that you will be able to transfer your innovation there :) !",
    "1302232": ""
  }
}