{
  "id": 168532,
  "title": "How to do k-fold validate in object detection?",
  "url": "/competitions/global-wheat-detection/discussion/168532",
  "author_name": "",
  "post_date": "2020-07-21T02:54:56.316984200Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>In other competition, we could use 5-fold validation to train 5 model and use the <strong>average</strong> of 5 model outputs of test imgaes as the final results. How to do it in object detection as the outputs are bounding box?</p>",
  "messages": [
    {
      "id": "937440",
      "postDate": "07/21/2020 02:54:56",
      "content": "<p>In other competition, we could use 5-fold validation to train 5 model and use the <strong>average</strong> of 5 model outputs of test imgaes as the final results. How to do it in object detection as the outputs are bounding box?</p>",
      "rawMarkdown": "In other competition, we could use 5-fold validation to train 5 model and use the **average** of 5 model outputs of test imgaes as the final results. How to do it in object detection as the outputs are bounding box?",
      "votes": null
    },
    {
      "id": "938510",
      "postDate": "07/21/2020 15:23:25",
      "content": "<p>You can use the ensemble_boxes library:\n<a href=\"https://www.kaggle.com/shonenkov/wbf-approach-for-ensemble\">https://www.kaggle.com/shonenkov/wbf-approach-for-ensemble</a></p>",
      "rawMarkdown": "You can use the ensemble_boxes library:\n[https://www.kaggle.com/shonenkov/wbf-approach-for-ensemble](https://www.kaggle.com/shonenkov/wbf-approach-for-ensemble)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 938510,
      "author_name": "jonykarki",
      "author_url": "",
      "post_date": "07/21/2020 15:23:25",
      "content": "<p>You can use the ensemble_boxes library:\n<a href=\"https://www.kaggle.com/shonenkov/wbf-approach-for-ensemble\">https://www.kaggle.com/shonenkov/wbf-approach-for-ensemble</a></p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "937440": "In other competition, we could use 5-fold validation to train 5 model and use the **average** of 5 model outputs of test imgaes as the final results. How to do it in object detection as the outputs are bounding box?",
    "938510": "You can use the ensemble_boxes library:\n[https://www.kaggle.com/shonenkov/wbf-approach-for-ensemble](https://www.kaggle.com/shonenkov/wbf-approach-for-ensemble)"
  },
  "source": "meta"
}