{
  "id": 307174,
  "title": "New to Kaggle",
  "url": "/competitions/tensorflow-great-barrier-reef/discussion/307174",
  "author_name": "",
  "post_date": "2022-02-13T03:57:29.653182200Z",
  "votes": 4,
  "comment_count": 6,
  "views": 0,
  "content": "<p>Hi everyone, I am a Kaggle beginner it would be great if anyone could guide me on what should I start with.</p>",
  "messages": [
    {
      "id": "1687678",
      "postDate": "02/13/2022 03:57:29",
      "content": "<p>Hi everyone, I am a Kaggle beginner it would be great if anyone could guide me on what should I start with.</p>",
      "rawMarkdown": "Hi everyone, I am a Kaggle beginner it would be great if anyone could guide me on what should I start with.",
      "votes": null
    },
    {
      "id": "1688359",
      "postDate": "02/13/2022 15:37:02",
      "content": "<p>When I first started kaggle,  i began with a baseline code and tuned it<br>\nI think this method is the quickest way to learn kaggle.<br>\nwith appropriate tuning of a baseline code, i could barely get the bronze in the second competition.</p>",
      "rawMarkdown": "When I first started kaggle,  i began with a baseline code and tuned it\nI think this method is the quickest way to learn kaggle.\nwith appropriate tuning of a baseline code, i could barely get the bronze in the second competition.",
      "votes": null
    },
    {
      "id": "1688362",
      "postDate": "02/13/2022 15:38:17",
      "content": "<p>Oh thanks for that but can give more details on it.</p>",
      "rawMarkdown": "Oh thanks for that but can give more details on it.",
      "votes": null
    },
    {
      "id": "1688377",
      "postDate": "02/13/2022 15:42:35",
      "content": "<p>Depends on how much you know about ML and programming. If you're a total newbie, consider learning more about Python first. Otherwise, I suggest taking a few public notebooks and playing around, so you can familiarize yourself with the many features of Kaggle.</p>\n<p>There's many different training notebooks which offer great advice. I learned a lot from others through public notebooks.</p>",
      "rawMarkdown": "Depends on how much you know about ML and programming. If you're a total newbie, consider learning more about Python first. Otherwise, I suggest taking a few public notebooks and playing around, so you can familiarize yourself with the many features of Kaggle.\n\nThere's many different training notebooks which offer great advice. I learned a lot from others through public notebooks.",
      "votes": null
    },
    {
      "id": "1688395",
      "postDate": "02/13/2022 15:47:22",
      "content": "<p>Sure. for example in this competition, you can find a lot of baseline notebooks<br>\nLike this one -&gt; <a href=\"https://www.kaggle.com/steamedsheep/yolov5-is-all-you-need\" target=\"_blank\">https://www.kaggle.com/steamedsheep/yolov5-is-all-you-need</a><br>\nIts notebook scored 0.579 using the yolov5s6 model with 3600 image size<br>\nwith the hint it can score around 0.57x , you can train the yolov5s model with different sizes like 3000,1280 and make a submission for it<br>\nor you can adjust augmentations of the yolov5 like rotation, hsv adjustment, and see the result<br>\nalso you can fix the baseline notebook with what you think is right. </p>\n<p>I mean you can start from someone else's product and renovate it</p>",
      "rawMarkdown": "Sure. for example in this competition, you can find a lot of baseline notebooks\nLike this one -> https://www.kaggle.com/steamedsheep/yolov5-is-all-you-need\nIts notebook scored 0.579 using the yolov5s6 model with 3600 image size\nwith the hint it can score around 0.57x , you can train the yolov5s model with different sizes like 3000,1280 and make a submission for it\nor you can adjust augmentations of the yolov5 like rotation, hsv adjustment, and see the result\nalso you can fix the baseline notebook with what you think is right. \n\nI mean you can start from someone else's product and renovate it",
      "votes": null
    },
    {
      "id": "1688431",
      "postDate": "02/13/2022 16:08:38",
      "content": "<p>Now I understood that will give it a try.</p>",
      "rawMarkdown": "Now I understood that will give it a try.",
      "votes": null
    },
    {
      "id": "1688437",
      "postDate": "02/13/2022 16:12:03",
      "content": "<p>Thanks will try that</p>",
      "rawMarkdown": "Thanks will try that",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1688359,
      "author_name": "deepkim",
      "author_url": "",
      "post_date": "02/13/2022 15:37:02",
      "content": "<p>When I first started kaggle,  i began with a baseline code and tuned it<br>\nI think this method is the quickest way to learn kaggle.<br>\nwith appropriate tuning of a baseline code, i could barely get the bronze in the second competition.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1688362,
          "author_name": "databro",
          "author_url": "",
          "post_date": "02/13/2022 15:38:17",
          "content": "<p>Oh thanks for that but can give more details on it.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1688395,
          "author_name": "deepkim",
          "author_url": "",
          "post_date": "02/13/2022 15:47:22",
          "content": "<p>Sure. for example in this competition, you can find a lot of baseline notebooks<br>\nLike this one -&gt; <a href=\"https://www.kaggle.com/steamedsheep/yolov5-is-all-you-need\" target=\"_blank\">https://www.kaggle.com/steamedsheep/yolov5-is-all-you-need</a><br>\nIts notebook scored 0.579 using the yolov5s6 model with 3600 image size<br>\nwith the hint it can score around 0.57x , you can train the yolov5s model with different sizes like 3000,1280 and make a submission for it<br>\nor you can adjust augmentations of the yolov5 like rotation, hsv adjustment, and see the result<br>\nalso you can fix the baseline notebook with what you think is right. </p>\n<p>I mean you can start from someone else's product and renovate it</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1688431,
          "author_name": "databro",
          "author_url": "",
          "post_date": "02/13/2022 16:08:38",
          "content": "<p>Now I understood that will give it a try.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1688377,
      "author_name": "liminchen1",
      "author_url": "",
      "post_date": "02/13/2022 15:42:35",
      "content": "<p>Depends on how much you know about ML and programming. If you're a total newbie, consider learning more about Python first. Otherwise, I suggest taking a few public notebooks and playing around, so you can familiarize yourself with the many features of Kaggle.</p>\n<p>There's many different training notebooks which offer great advice. I learned a lot from others through public notebooks.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1688437,
          "author_name": "databro",
          "author_url": "",
          "post_date": "02/13/2022 16:12:03",
          "content": "<p>Thanks will try that</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1687678": "Hi everyone, I am a Kaggle beginner it would be great if anyone could guide me on what should I start with.",
    "1688359": "When I first started kaggle,  i began with a baseline code and tuned it\nI think this method is the quickest way to learn kaggle.\nwith appropriate tuning of a baseline code, i could barely get the bronze in the second competition.",
    "1688362": "Oh thanks for that but can give more details on it.",
    "1688377": "Depends on how much you know about ML and programming. If you're a total newbie, consider learning more about Python first. Otherwise, I suggest taking a few public notebooks and playing around, so you can familiarize yourself with the many features of Kaggle.\n\nThere's many different training notebooks which offer great advice. I learned a lot from others through public notebooks.",
    "1688395": "Sure. for example in this competition, you can find a lot of baseline notebooks\nLike this one -> https://www.kaggle.com/steamedsheep/yolov5-is-all-you-need\nIts notebook scored 0.579 using the yolov5s6 model with 3600 image size\nwith the hint it can score around 0.57x , you can train the yolov5s model with different sizes like 3000,1280 and make a submission for it\nor you can adjust augmentations of the yolov5 like rotation, hsv adjustment, and see the result\nalso you can fix the baseline notebook with what you think is right. \n\nI mean you can start from someone else's product and renovate it",
    "1688431": "Now I understood that will give it a try.",
    "1688437": "Thanks will try that"
  },
  "source": "meta"
}