{
  "id": 480172,
  "title": "Best starting point?",
  "url": "/competitions/hms-harmful-brain-activity-classification/discussion/480172",
  "author_name": "",
  "post_date": "2024-02-27T14:07:00.114417900Z",
  "votes": 4,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Hi,  i was trying the Home Credit comp, but since there will be a break, I decided to try this competition for the meantime. Since I'm late, I wanted to know what the best starting point is to catch up in terms of notebooks.</p>",
  "messages": [
    {
      "id": "2671439",
      "postDate": "02/27/2024 14:07:00",
      "content": "<p>Hi,  i was trying the Home Credit comp, but since there will be a break, I decided to try this competition for the meantime. Since I'm late, I wanted to know what the best starting point is to catch up in terms of notebooks.</p>",
      "rawMarkdown": "Hi,  i was trying the Home Credit comp, but since there will be a break, I decided to try this competition for the meantime. Since I'm late, I wanted to know what the best starting point is to catch up in terms of notebooks.",
      "votes": null
    },
    {
      "id": "2673321",
      "postDate": "02/28/2024 16:14:05",
      "content": "<p>From my point of view, the easiest way to start is to focus on CNN notebooks using keras that use spectograms offered by the competition…</p>\n<p>Then you can focus on studying the creation of custom spectograms and trying to improve your model.</p>",
      "rawMarkdown": "From my point of view, the easiest way to start is to focus on CNN notebooks using keras that use spectograms offered by the competition...\n\nThen you can focus on studying the creation of custom spectograms and trying to improve your model.",
      "votes": null
    },
    {
      "id": "2674681",
      "postDate": "02/29/2024 13:42:38",
      "content": "<p><a href=\"https://www.kaggle.com/manuelandersen\" target=\"_blank\">@manuelandersen</a> I recommend checking out the <a href=\"https://www.kaggle.com/code/nartaa/features-head-starter\" target=\"_blank\">Features+Head Starter</a> notebook. It offers all-in-one solution allowing you to train individual models, conduct ensemble, and make submissions, all within a single notebook. Additionally, it comes with pre-trained models included, providing a strong foundation with a leaderboard score of 0.34.</p>",
      "rawMarkdown": "manuelandersen I recommend checking out the [Features+Head Starter](https://www.kaggle.com/code/nartaa/features-head-starter) notebook. It offers all-in-one solution allowing you to train individual models, conduct ensemble, and make submissions, all within a single notebook. Additionally, it comes with pre-trained models included, providing a strong foundation with a leaderboard score of 0.34.",
      "votes": null
    },
    {
      "id": "2674712",
      "postDate": "02/29/2024 14:05:07",
      "content": "<p>thks for the tips <a href=\"https://www.kaggle.com/rafaelzimmermann1\" target=\"_blank\">@rafaelzimmermann1</a>!</p>",
      "rawMarkdown": "thks for the tips @rafaelzimmermann1!",
      "votes": null
    },
    {
      "id": "2674721",
      "postDate": "02/29/2024 14:06:50",
      "content": "<p>thks <a href=\"https://www.kaggle.com/nartaa\" target=\"_blank\">@nartaa</a>! i will take a look into the notebook</p>",
      "rawMarkdown": "thks @nartaa! i will take a look into the notebook",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2673321,
      "author_name": "rafaelzimmermann1",
      "author_url": "",
      "post_date": "02/28/2024 16:14:05",
      "content": "<p>From my point of view, the easiest way to start is to focus on CNN notebooks using keras that use spectograms offered by the competition…</p>\n<p>Then you can focus on studying the creation of custom spectograms and trying to improve your model.</p>",
      "votes": null,
      "replies": [
        {
          "id": 2674712,
          "author_name": "manuelandersen",
          "author_url": "",
          "post_date": "02/29/2024 14:05:07",
          "content": "<p>thks for the tips <a href=\"https://www.kaggle.com/rafaelzimmermann1\" target=\"_blank\">@rafaelzimmermann1</a>!</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 2674681,
      "author_name": "nartaa",
      "author_url": "",
      "post_date": "02/29/2024 13:42:38",
      "content": "<p><a href=\"https://www.kaggle.com/manuelandersen\" target=\"_blank\">@manuelandersen</a> I recommend checking out the <a href=\"https://www.kaggle.com/code/nartaa/features-head-starter\" target=\"_blank\">Features+Head Starter</a> notebook. It offers all-in-one solution allowing you to train individual models, conduct ensemble, and make submissions, all within a single notebook. Additionally, it comes with pre-trained models included, providing a strong foundation with a leaderboard score of 0.34.</p>",
      "votes": null,
      "replies": [
        {
          "id": 2674721,
          "author_name": "manuelandersen",
          "author_url": "",
          "post_date": "02/29/2024 14:06:50",
          "content": "<p>thks <a href=\"https://www.kaggle.com/nartaa\" target=\"_blank\">@nartaa</a>! i will take a look into the notebook</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2671439": "Hi,  i was trying the Home Credit comp, but since there will be a break, I decided to try this competition for the meantime. Since I'm late, I wanted to know what the best starting point is to catch up in terms of notebooks.",
    "2673321": "From my point of view, the easiest way to start is to focus on CNN notebooks using keras that use spectograms offered by the competition...\n\nThen you can focus on studying the creation of custom spectograms and trying to improve your model.",
    "2674681": "manuelandersen I recommend checking out the [Features+Head Starter](https://www.kaggle.com/code/nartaa/features-head-starter) notebook. It offers all-in-one solution allowing you to train individual models, conduct ensemble, and make submissions, all within a single notebook. Additionally, it comes with pre-trained models included, providing a strong foundation with a leaderboard score of 0.34.",
    "2674712": "thks for the tips @rafaelzimmermann1!",
    "2674721": "thks @nartaa! i will take a look into the notebook"
  },
  "source": "meta"
}