{
  "id": 615030,
  "title": "What do you think th final score could be? Share your thoughts!",
  "url": "/competitions/csiro-biomass/discussion/615030",
  "author_name": "",
  "post_date": "2025-11-08T10:50:16.791749500Z",
  "votes": 2,
  "comment_count": 9,
  "views": 0,
  "content": "<p>What do you think th final score could be? Share your thoughts!</p>",
  "messages": [
    {
      "id": "3312964",
      "postDate": "11/08/2025 10:50:16",
      "content": "<p>What do you think th final score could be? Share your thoughts!</p>",
      "rawMarkdown": "What do you think th final score could be? Share your thoughts!",
      "votes": null
    },
    {
      "id": "3312979",
      "postDate": "11/08/2025 11:24:27",
      "content": "<p>Would be nice to get above 0.8… </p>",
      "rawMarkdown": "Would be nice to get above 0.8...",
      "votes": null
    },
    {
      "id": "3313089",
      "postDate": "11/08/2025 16:21:20",
      "content": "<p>I feel the top score would be ~0.85</p>",
      "rawMarkdown": "I feel the top score would be ~0.85",
      "votes": null
    },
    {
      "id": "3313111",
      "postDate": "11/08/2025 17:04:01",
      "content": "<p>I’m wondering, what score would be considered \"good enough\" so that our solution would be useful in real-world industry applications?</p>\n<p>It would be great if we could achieve that</p>",
      "rawMarkdown": "I’m wondering, what score would be considered \"good enough\" so that our solution would be useful in real-world industry applications?\n\nIt would be great if we could achieve that",
      "votes": null
    },
    {
      "id": "3313190",
      "postDate": "11/08/2025 21:40:05",
      "content": "<p>In a real world situation there are several pieces of information that are likely available that we do not have: camera focus depth and aperture would allow for the model to predict height through more physically relevant means, for example. Also, I imagine that this would be a useful thing for a farmer to trend over time, so outliers are far less likely as the trend would also be an input into that model. </p>",
      "rawMarkdown": "In a real world situation there are several pieces of information that are likely available that we do not have: camera focus depth and aperture would allow for the model to predict height through more physically relevant means, for example. Also, I imagine that this would be a useful thing for a farmer to trend over time, so outliers are far less likely as the trend would also be an input into that model.",
      "votes": null
    },
    {
      "id": "3313732",
      "postDate": "11/09/2025 17:55:04",
      "content": "<p>How much time does it take for you guys to train , if i may ask.\nAsking because i will for sure burn my 30hr/week in 2 days on Kaggle.\nAnd a 16GB VRAM GPU is too slow for the model i would like to train on. Do you guys use other GPU providers or something ?</p>",
      "rawMarkdown": "How much time does it take for you guys to train , if i may ask.\nAsking because i will for sure burn my 30hr/week in 2 days on Kaggle.\nAnd a 16GB VRAM GPU is too slow for the model i would like to train on. Do you guys use other GPU providers or something ?",
      "votes": null
    },
    {
      "id": "3313739",
      "postDate": "11/09/2025 18:09:16",
      "content": "<p>It should surely go over 0.8</p>",
      "rawMarkdown": "It should surely go over 0.8",
      "votes": null
    },
    {
      "id": "3313748",
      "postDate": "11/09/2025 18:33:43",
      "content": "<p>I am doing all of my training at home (RTX 3070, Ryzen 5600) or on Kaggle at the moment.  </p>",
      "rawMarkdown": "I am doing all of my training at home (RTX 3070, Ryzen 5600) or on Kaggle at the moment.",
      "votes": null
    },
    {
      "id": "3313749",
      "postDate": "11/09/2025 18:36:37",
      "content": "<p>40 min on kaggle t4*2</p>",
      "rawMarkdown": "40 min on kaggle t4*2",
      "votes": null
    },
    {
      "id": "3315574",
      "postDate": "11/10/2025 17:05:02",
      "content": "<p>Top score will be 0.75 or less.  The images are low quality with lots of noise.  The sample size is small.  </p>",
      "rawMarkdown": "Top score will be 0.75 or less.  The images are low quality with lots of noise.  The sample size is small.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3312979,
      "author_name": "timkrige",
      "author_url": "",
      "post_date": "11/08/2025 11:24:27",
      "content": "<p>Would be nice to get above 0.8… </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 3313089,
      "author_name": "paritoshtripathi5",
      "author_url": "",
      "post_date": "11/08/2025 16:21:20",
      "content": "<p>I feel the top score would be ~0.85</p>",
      "votes": null,
      "replies": [
        {
          "id": 3313732,
          "author_name": "skathaitrooney",
          "author_url": "",
          "post_date": "11/09/2025 17:55:04",
          "content": "<p>How much time does it take for you guys to train , if i may ask.\nAsking because i will for sure burn my 30hr/week in 2 days on Kaggle.\nAnd a 16GB VRAM GPU is too slow for the model i would like to train on. Do you guys use other GPU providers or something ?</p>",
          "votes": null,
          "replies": [
            {
              "id": 3313748,
              "author_name": "timkrige",
              "author_url": "",
              "post_date": "11/09/2025 18:33:43",
              "content": "<p>I am doing all of my training at home (RTX 3070, Ryzen 5600) or on Kaggle at the moment.  </p>",
              "votes": null,
              "replies": []
            },
            {
              "id": 3313749,
              "author_name": "paritoshtripathi5",
              "author_url": "",
              "post_date": "11/09/2025 18:36:37",
              "content": "<p>40 min on kaggle t4*2</p>",
              "votes": null,
              "replies": []
            }
          ]
        }
      ]
    },
    {
      "id": 3313111,
      "author_name": "mushfirat",
      "author_url": "",
      "post_date": "11/08/2025 17:04:01",
      "content": "<p>I’m wondering, what score would be considered \"good enough\" so that our solution would be useful in real-world industry applications?</p>\n<p>It would be great if we could achieve that</p>",
      "votes": null,
      "replies": [
        {
          "id": 3313190,
          "author_name": "timkrige",
          "author_url": "",
          "post_date": "11/08/2025 21:40:05",
          "content": "<p>In a real world situation there are several pieces of information that are likely available that we do not have: camera focus depth and aperture would allow for the model to predict height through more physically relevant means, for example. Also, I imagine that this would be a useful thing for a farmer to trend over time, so outliers are far less likely as the trend would also be an input into that model. </p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 3313739,
      "author_name": "cid007",
      "author_url": "",
      "post_date": "11/09/2025 18:09:16",
      "content": "<p>It should surely go over 0.8</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 3315574,
      "author_name": "pcjimmmy",
      "author_url": "",
      "post_date": "11/10/2025 17:05:02",
      "content": "<p>Top score will be 0.75 or less.  The images are low quality with lots of noise.  The sample size is small.  </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3312964": "What do you think th final score could be? Share your thoughts!",
    "3312979": "Would be nice to get above 0.8...",
    "3313089": "I feel the top score would be ~0.85",
    "3313111": "I’m wondering, what score would be considered \"good enough\" so that our solution would be useful in real-world industry applications?\n\nIt would be great if we could achieve that",
    "3313190": "In a real world situation there are several pieces of information that are likely available that we do not have: camera focus depth and aperture would allow for the model to predict height through more physically relevant means, for example. Also, I imagine that this would be a useful thing for a farmer to trend over time, so outliers are far less likely as the trend would also be an input into that model.",
    "3313732": "How much time does it take for you guys to train , if i may ask.\nAsking because i will for sure burn my 30hr/week in 2 days on Kaggle.\nAnd a 16GB VRAM GPU is too slow for the model i would like to train on. Do you guys use other GPU providers or something ?",
    "3313739": "It should surely go over 0.8",
    "3313748": "I am doing all of my training at home (RTX 3070, Ryzen 5600) or on Kaggle at the moment.",
    "3313749": "40 min on kaggle t4*2",
    "3315574": "Top score will be 0.75 or less.  The images are low quality with lots of noise.  The sample size is small."
  },
  "source": "meta"
}