{
  "id": 663876,
  "title": "Does anyone use models from the Efficnet series? What is the highest score achieved?",
  "url": "/competitions/csiro-biomass/discussion/663876",
  "author_name": "",
  "post_date": "2025-12-21T07:16:01.993581400Z",
  "votes": null,
  "comment_count": 2,
  "views": 0,
  "content": "<p>The best LB score I got using efficient-b2 is 0.59. After switching to B4, the score didn't increase. Is everyone using large models? Training large models is too slow, but it seems they bring very good results.</p>",
  "messages": [
    {
      "id": "3379919",
      "postDate": "12/21/2025 07:16:01",
      "content": "<p>The best LB score I got using efficient-b2 is 0.59. After switching to B4, the score didn't increase. Is everyone using large models? Training large models is too slow, but it seems they bring very good results.</p>",
      "rawMarkdown": "The best LB score I got using efficient-b2 is 0.59. After switching to B4, the score didn't increase. Is everyone using large models? Training large models is too slow, but it seems they bring very good results.",
      "votes": null
    },
    {
      "id": "3379954",
      "postDate": "12/21/2025 09:10:00",
      "content": "<p>The same thing, my maximum score of 0.59 drops sharply with any changes.</p>",
      "rawMarkdown": "The same thing, my maximum score of 0.59 drops sharply with any changes.",
      "votes": null
    },
    {
      "id": "3379979",
      "postDate": "12/21/2025 11:00:48",
      "content": "<p>I got 0.6 and 0.59 using 2 different variants. I haven't done any serious parameter optimization so far, just plugged them into my existing pipeline.</p>",
      "rawMarkdown": "I got 0.6 and 0.59 using 2 different variants. I haven't done any serious parameter optimization so far, just plugged them into my existing pipeline.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3379954,
      "author_name": "markdjadchenko",
      "author_url": "",
      "post_date": "12/21/2025 09:10:00",
      "content": "<p>The same thing, my maximum score of 0.59 drops sharply with any changes.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 3379979,
      "author_name": "mohib94",
      "author_url": "",
      "post_date": "12/21/2025 11:00:48",
      "content": "<p>I got 0.6 and 0.59 using 2 different variants. I haven't done any serious parameter optimization so far, just plugged them into my existing pipeline.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3379919": "The best LB score I got using efficient-b2 is 0.59. After switching to B4, the score didn't increase. Is everyone using large models? Training large models is too slow, but it seems they bring very good results.",
    "3379954": "The same thing, my maximum score of 0.59 drops sharply with any changes.",
    "3379979": "I got 0.6 and 0.59 using 2 different variants. I haven't done any serious parameter optimization so far, just plugged them into my existing pipeline."
  },
  "source": "meta"
}