{
  "id": 269609,
  "title": "🔥Wow four people with score more than 0.87🔥",
  "url": "/competitions/rsna-miccai-brain-tumor-radiogenomic-classification/discussion/269609",
  "author_name": "",
  "post_date": "2021-09-01T11:05:45.060118800Z",
  "votes": 3,
  "comment_count": 5,
  "views": 0,
  "content": "<p>Hey, Do you have any opinion for increase our score more than 0.8 ?<br>\nI tried with different 3D model and used transfer learning but i didn't get more than 0.66 score.</p>",
  "messages": [
    {
      "id": "1498966",
      "postDate": "09/01/2021 11:05:45",
      "content": "<p>Hey, Do you have any opinion for increase our score more than 0.8 ?<br>\nI tried with different 3D model and used transfer learning but i didn't get more than 0.66 score.</p>",
      "rawMarkdown": "Hey, Do you have any opinion for increase our score more than 0.8 ?\nI tried with different 3D model and used transfer learning but i didn't get more than 0.66 score.",
      "votes": null
    },
    {
      "id": "1500106",
      "postDate": "09/02/2021 06:51:21",
      "content": "<p>Most of the models are just overfitting, you should check additional scores like accuracy too. </p>",
      "rawMarkdown": "Most of the models are just overfitting, you should check additional scores like accuracy too.",
      "votes": null
    },
    {
      "id": "1500242",
      "postDate": "09/02/2021 08:27:09",
      "content": "<p>transfer learning is useful or not?<br>\nresnet, vgg, densenet model</p>",
      "rawMarkdown": "transfer learning is useful or not?\nresnet, vgg, densenet model",
      "votes": null
    },
    {
      "id": "1500348",
      "postDate": "09/02/2021 10:03:28",
      "content": "<p>Not useful currently, it is not a model centric problem. I think it is more of a problem with the data and the correct normalisation of the 3d data</p>",
      "rawMarkdown": "Not useful currently, it is not a model centric problem. I think it is more of a problem with the data and the correct normalisation of the 3d data",
      "votes": null
    },
    {
      "id": "1500359",
      "postDate": "09/02/2021 10:08:15",
      "content": "<p>Do you know simple 3D models in keras? i found one model from this site that many people used in this competition.</p>",
      "rawMarkdown": "Do you know simple 3D models in keras? i found one model from this site that many people used in this competition.",
      "votes": null
    },
    {
      "id": "1500380",
      "postDate": "09/02/2021 10:39:56",
      "content": "<p>Keras itself has a very good and easy 3dcnn in their documentation <a href=\"url\" target=\"_blank\">https://keras.io/examples/vision/3D_image_classification/</a></p>",
      "rawMarkdown": "Keras itself has a very good and easy 3dcnn in their documentation [https://keras.io/examples/vision/3D_image_classification/](url)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1500106,
      "author_name": "skyyy93",
      "author_url": "",
      "post_date": "09/02/2021 06:51:21",
      "content": "<p>Most of the models are just overfitting, you should check additional scores like accuracy too. </p>",
      "votes": null,
      "replies": [
        {
          "id": 1500242,
          "author_name": "mohammadhosein1998",
          "author_url": "",
          "post_date": "09/02/2021 08:27:09",
          "content": "<p>transfer learning is useful or not?<br>\nresnet, vgg, densenet model</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1500348,
          "author_name": "skyyy93",
          "author_url": "",
          "post_date": "09/02/2021 10:03:28",
          "content": "<p>Not useful currently, it is not a model centric problem. I think it is more of a problem with the data and the correct normalisation of the 3d data</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1500359,
          "author_name": "mohammadhosein1998",
          "author_url": "",
          "post_date": "09/02/2021 10:08:15",
          "content": "<p>Do you know simple 3D models in keras? i found one model from this site that many people used in this competition.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1500380,
          "author_name": "skyyy93",
          "author_url": "",
          "post_date": "09/02/2021 10:39:56",
          "content": "<p>Keras itself has a very good and easy 3dcnn in their documentation <a href=\"url\" target=\"_blank\">https://keras.io/examples/vision/3D_image_classification/</a></p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1498966": "Hey, Do you have any opinion for increase our score more than 0.8 ?\nI tried with different 3D model and used transfer learning but i didn't get more than 0.66 score.",
    "1500106": "Most of the models are just overfitting, you should check additional scores like accuracy too.",
    "1500242": "transfer learning is useful or not?\nresnet, vgg, densenet model",
    "1500348": "Not useful currently, it is not a model centric problem. I think it is more of a problem with the data and the correct normalisation of the 3d data",
    "1500359": "Do you know simple 3D models in keras? i found one model from this site that many people used in this competition.",
    "1500380": "Keras itself has a very good and easy 3dcnn in their documentation [https://keras.io/examples/vision/3D_image_classification/](url)"
  },
  "source": "meta"
}