{
  "id": 296788,
  "title": "Question to the guys at top of the leaderboard",
  "url": "/competitions/tensorflow-great-barrier-reef/discussion/296788",
  "author_name": "",
  "post_date": "2021-12-23T15:30:51.401015200Z",
  "votes": 14,
  "comment_count": 5,
  "views": 0,
  "content": "<p>I hope this question is allowed so here we go. (It may even help out other people like me who are new at this)</p>\n<p>I started Kaggle relatively recently (its only my 2nd competition that I am actively trying). <br>\nSo this is kind of noob question but what are your platforms of choice for training models?</p>\n<p>I understand there are 3 options, <br>\n-use Kaggle's own platform<br>\n-google collab/pro/pro+<br>\n-use our own pc</p>\n<p>Currently I don't have pro/pro+ subscription on Collab, nor do I have PC with powerful GPU to train on my own. My only option is Kaggle's platform atm.</p>\n<p>I am happy with the service I am getting here on kaggle but my final question or curiosity is<br>\n-Is it possible to be in the top 20 of leaderboard with 35-40 hrs of gpu per week?<br>\n-Are any of you in the top solely using Kaggle's free GPU quota? <br>\n-If not, is Collab pro (9.99 per month) enough or do I need pro+?</p>\n<p><strong>Any insights and helpful comments are appreciated. Thank you so much.</strong><br>\n(I understand I can just google it but getting answers from someone actively participating will be much more insightful than going through some random websites.)</p>",
  "messages": [
    {
      "id": "1627177",
      "postDate": "12/23/2021 15:30:51",
      "content": "<p>I hope this question is allowed so here we go. (It may even help out other people like me who are new at this)</p>\n<p>I started Kaggle relatively recently (its only my 2nd competition that I am actively trying). <br>\nSo this is kind of noob question but what are your platforms of choice for training models?</p>\n<p>I understand there are 3 options, <br>\n-use Kaggle's own platform<br>\n-google collab/pro/pro+<br>\n-use our own pc</p>\n<p>Currently I don't have pro/pro+ subscription on Collab, nor do I have PC with powerful GPU to train on my own. My only option is Kaggle's platform atm.</p>\n<p>I am happy with the service I am getting here on kaggle but my final question or curiosity is<br>\n-Is it possible to be in the top 20 of leaderboard with 35-40 hrs of gpu per week?<br>\n-Are any of you in the top solely using Kaggle's free GPU quota? <br>\n-If not, is Collab pro (9.99 per month) enough or do I need pro+?</p>\n<p><strong>Any insights and helpful comments are appreciated. Thank you so much.</strong><br>\n(I understand I can just google it but getting answers from someone actively participating will be much more insightful than going through some random websites.)</p>",
      "rawMarkdown": "I hope this question is allowed so here we go. (It may even help out other people like me who are new at this)\n\nI started Kaggle relatively recently (its only my 2nd competition that I am actively trying). \nSo this is kind of noob question but what are your platforms of choice for training models?\n\nI understand there are 3 options, \n-use Kaggle's own platform\n-google collab/pro/pro+\n-use our own pc\n\nCurrently I don't have pro/pro+ subscription on Collab, nor do I have PC with powerful GPU to train on my own. My only option is Kaggle's platform atm.\n\nI am happy with the service I am getting here on kaggle but my final question or curiosity is\n-Is it possible to be in the top 20 of leaderboard with 35-40 hrs of gpu per week?\n-Are any of you in the top solely using Kaggle's free GPU quota? \n-If not, is Collab pro (9.99 per month) enough or do I need pro+?\n\n**Any insights and helpful comments are appreciated. Thank you so much.**\n(I understand I can just google it but getting answers from someone actively participating will be much more insightful than going through some random websites.)",
      "votes": null
    },
    {
      "id": "1627263",
      "postDate": "12/23/2021 17:35:48",
      "content": "<p>My answer to all three of your questions would be: yes, yes, yes :-) (although I am not in the top20)</p>",
      "rawMarkdown": "My answer to all three of your questions would be: yes, yes, yes :-) (although I am not in the top20)",
      "votes": null
    },
    {
      "id": "1627325",
      "postDate": "12/23/2021 18:54:20",
      "content": "<p>oh wow 22nd using only kaggle, that gives me so much hope tysm.</p>",
      "rawMarkdown": "oh wow 22nd using only kaggle, that gives me so much hope tysm.",
      "votes": null
    },
    {
      "id": "1627388",
      "postDate": "12/23/2021 20:05:26",
      "content": "<p>Our current score/position, 13th, is from just free Kaggle. I do have Colab Pro though and I have moved to that for the bigger longer running models. </p>",
      "rawMarkdown": "Our current score/position, 13th, is from just free Kaggle. I do have Colab Pro though and I have moved to that for the bigger longer running models.",
      "votes": null
    },
    {
      "id": "1627594",
      "postDate": "12/24/2021 03:32:00",
      "content": "<p>That's amazing, thank you so much for sharing. </p>",
      "rawMarkdown": "That's amazing, thank you so much for sharing.",
      "votes": null
    },
    {
      "id": "1650424",
      "postDate": "01/15/2022 03:29:27",
      "content": "<p>Thank you for your insights</p>",
      "rawMarkdown": "Thank you for your insights",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1627263,
      "author_name": "hannes82",
      "author_url": "",
      "post_date": "12/23/2021 17:35:48",
      "content": "<p>My answer to all three of your questions would be: yes, yes, yes :-) (although I am not in the top20)</p>",
      "votes": null,
      "replies": [
        {
          "id": 1627325,
          "author_name": "keshhere",
          "author_url": "",
          "post_date": "12/23/2021 18:54:20",
          "content": "<p>oh wow 22nd using only kaggle, that gives me so much hope tysm.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1627388,
      "author_name": "mutantspore",
      "author_url": "",
      "post_date": "12/23/2021 20:05:26",
      "content": "<p>Our current score/position, 13th, is from just free Kaggle. I do have Colab Pro though and I have moved to that for the bigger longer running models. </p>",
      "votes": null,
      "replies": [
        {
          "id": 1627594,
          "author_name": "keshhere",
          "author_url": "",
          "post_date": "12/24/2021 03:32:00",
          "content": "<p>That's amazing, thank you so much for sharing. </p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1650424,
      "author_name": "nyanswanaung",
      "author_url": "",
      "post_date": "01/15/2022 03:29:27",
      "content": "<p>Thank you for your insights</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1627177": "I hope this question is allowed so here we go. (It may even help out other people like me who are new at this)\n\nI started Kaggle relatively recently (its only my 2nd competition that I am actively trying). \nSo this is kind of noob question but what are your platforms of choice for training models?\n\nI understand there are 3 options, \n-use Kaggle's own platform\n-google collab/pro/pro+\n-use our own pc\n\nCurrently I don't have pro/pro+ subscription on Collab, nor do I have PC with powerful GPU to train on my own. My only option is Kaggle's platform atm.\n\nI am happy with the service I am getting here on kaggle but my final question or curiosity is\n-Is it possible to be in the top 20 of leaderboard with 35-40 hrs of gpu per week?\n-Are any of you in the top solely using Kaggle's free GPU quota? \n-If not, is Collab pro (9.99 per month) enough or do I need pro+?\n\n**Any insights and helpful comments are appreciated. Thank you so much.**\n(I understand I can just google it but getting answers from someone actively participating will be much more insightful than going through some random websites.)",
    "1627263": "My answer to all three of your questions would be: yes, yes, yes :-) (although I am not in the top20)",
    "1627325": "oh wow 22nd using only kaggle, that gives me so much hope tysm.",
    "1627388": "Our current score/position, 13th, is from just free Kaggle. I do have Colab Pro though and I have moved to that for the bigger longer running models.",
    "1627594": "That's amazing, thank you so much for sharing.",
    "1650424": "Thank you for your insights"
  },
  "source": "meta"
}