{
  "id": 570344,
  "title": "How can a single YOLOV8 model train scores above LB0.62?",
  "url": "/competitions/byu-locating-bacterial-flagellar-motors-2025/discussion/570344",
  "author_name": "",
  "post_date": "2025-03-27T07:52:53.885686800Z",
  "votes": 7,
  "comment_count": 2,
  "views": 0,
  "content": "<p>How can a single YOLOV8 model train scores above LB0.62?</p>",
  "messages": [
    {
      "id": "3160839",
      "postDate": "03/27/2025 07:52:53",
      "content": "<p>How can a single YOLOV8 model train scores above LB0.62?</p>",
      "rawMarkdown": "How can a single YOLOV8 model train scores above LB0.62?",
      "votes": null
    },
    {
      "id": "3161925",
      "postDate": "03/28/2025 15:37:21",
      "content": "<p>Improved postprocessing techniques, tuned hyperparameters, better augmentations, threshold tuning, dataset preprocessing techniques, model ensembling, variations of YOLOv8, and so on.  There are a lot of techniques to get a few better points.</p>",
      "rawMarkdown": "Improved postprocessing techniques, tuned hyperparameters, better augmentations, threshold tuning, dataset preprocessing techniques, model ensembling, variations of YOLOv8, and so on.  There are a lot of techniques to get a few better points.",
      "votes": null
    },
    {
      "id": "3178628",
      "postDate": "04/14/2025 12:04:06",
      "content": "<p>try experimenting with various hyperparameters, and apply some preprocessing</p>",
      "rawMarkdown": "try experimenting with various hyperparameters, and apply some preprocessing",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3161925,
      "author_name": "connorjd",
      "author_url": "",
      "post_date": "03/28/2025 15:37:21",
      "content": "<p>Improved postprocessing techniques, tuned hyperparameters, better augmentations, threshold tuning, dataset preprocessing techniques, model ensembling, variations of YOLOv8, and so on.  There are a lot of techniques to get a few better points.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 3178628,
      "author_name": "aksh1t",
      "author_url": "",
      "post_date": "04/14/2025 12:04:06",
      "content": "<p>try experimenting with various hyperparameters, and apply some preprocessing</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3160839": "How can a single YOLOV8 model train scores above LB0.62?",
    "3161925": "Improved postprocessing techniques, tuned hyperparameters, better augmentations, threshold tuning, dataset preprocessing techniques, model ensembling, variations of YOLOv8, and so on.  There are a lot of techniques to get a few better points.",
    "3178628": "try experimenting with various hyperparameters, and apply some preprocessing"
  },
  "source": "meta"
}