{
  "id": 251977,
  "title": "Please tell me about the rules of the code competition",
  "url": "/competitions/mlb-player-digital-engagement-forecasting/discussion/251977",
  "author_name": "",
  "post_date": "2021-07-09T21:32:05.350236100Z",
  "votes": null,
  "comment_count": 5,
  "views": 0,
  "content": "<p>Hi,</p>\n<p>This is my first time participating in a code competition, and I'm not sure if I understand the rules correctly, so please let me know.</p>\n<p>In the rules section of the competition, there is a statement that says:</p>\n<blockquote>\n  <p>CPU Notebook &lt;= 6 hours run-time<br>\n  GPU Notebook &lt;= 6 hours run-time<br>\n  Internet access disabled<br>\n  Freely &amp; publicly available external data is allowed, including pre-trained models<br>\n  Submissions must be made using the&nbsp;mlb&nbsp;python module</p>\n</blockquote>\n<p>Is this time limit required only for inference?</p>\n<p>In other words, if I spend a long time (more than 6 hours) training a model locally (or in an environment other than KaggleNotebook), upload the trained model to Kaggle as a PrivateDataset, and then use that model to perform inference in less than 6 hours on KaggleNotebook, does this satisfy the rule?</p>\n<p>Let me know if I should ask in a different place.<br>\nI’m sorry if a similar topic has already been posted.</p>\n<p>Thanks in advance,</p>",
  "messages": [
    {
      "id": "1382445",
      "postDate": "07/09/2021 21:32:05",
      "content": "<p>Hi,</p>\n<p>This is my first time participating in a code competition, and I'm not sure if I understand the rules correctly, so please let me know.</p>\n<p>In the rules section of the competition, there is a statement that says:</p>\n<blockquote>\n  <p>CPU Notebook &lt;= 6 hours run-time<br>\n  GPU Notebook &lt;= 6 hours run-time<br>\n  Internet access disabled<br>\n  Freely &amp; publicly available external data is allowed, including pre-trained models<br>\n  Submissions must be made using the&nbsp;mlb&nbsp;python module</p>\n</blockquote>\n<p>Is this time limit required only for inference?</p>\n<p>In other words, if I spend a long time (more than 6 hours) training a model locally (or in an environment other than KaggleNotebook), upload the trained model to Kaggle as a PrivateDataset, and then use that model to perform inference in less than 6 hours on KaggleNotebook, does this satisfy the rule?</p>\n<p>Let me know if I should ask in a different place.<br>\nI’m sorry if a similar topic has already been posted.</p>\n<p>Thanks in advance,</p>",
      "rawMarkdown": "Hi,\n\nThis is my first time participating in a code competition, and I'm not sure if I understand the rules correctly, so please let me know.\n\nIn the rules section of the competition, there is a statement that says:\n\n> CPU Notebook <= 6 hours run-time\n> GPU Notebook <= 6 hours run-time\n> Internet access disabled\n> Freely & publicly available external data is allowed, including pre-trained models\n> Submissions must be made using the mlb python module\n\nIs this time limit required only for inference?\n\nIn other words, if I spend a long time (more than 6 hours) training a model locally (or in an environment other than KaggleNotebook), upload the trained model to Kaggle as a PrivateDataset, and then use that model to perform inference in less than 6 hours on KaggleNotebook, does this satisfy the rule?\n\nLet me know if I should ask in a different place.\nI’m sorry if a similar topic has already been posted.\n\nThanks in advance,",
      "votes": null
    },
    {
      "id": "1382487",
      "postDate": "07/09/2021 23:20:02",
      "content": "<p>Time limit is for the last submission - for the inference.    You can spend a decade off line training a model.</p>\n<p>And yes - this topic has been asked and answered in almost every competition I have been in for the past years.</p>",
      "rawMarkdown": "Time limit is for the last submission - for the inference.    You can spend a decade off line training a model.\n\nAnd yes - this topic has been asked and answered in almost every competition I have been in for the past years.",
      "votes": null
    },
    {
      "id": "1382647",
      "postDate": "07/10/2021 05:23:39",
      "content": "<p>I understood.<br>\nThank you very much!</p>",
      "rawMarkdown": "I understood.\nThank you very much!",
      "votes": null
    },
    {
      "id": "1382657",
      "postDate": "07/10/2021 05:41:19",
      "content": "<p>Your doing well so far for your first Kaggle - been here for several years.  Pretty much always a helpful group of people who participate in these competitions.  Made many trips to Japan over the years to visit at Toyota and Honda - never got north to see your part of Japan.  It's a beautiful country.</p>",
      "rawMarkdown": "Your doing well so far for your first Kaggle - been here for several years.  Pretty much always a helpful group of people who participate in these competitions.  Made many trips to Japan over the years to visit at Toyota and Honda - never got north to see your part of Japan.  It's a beautiful country.",
      "votes": null
    },
    {
      "id": "1402033",
      "postDate": "07/27/2021 18:41:14",
      "content": "<p>Hey Jelly, I have a similar concern. Can we upload our models as a private dataset? Asking because my submission threw a 'Notebook Exceeded Allowed Compute<br>\nError' when I used one. </p>",
      "rawMarkdown": "Hey Jelly, I have a similar concern. Can we upload our models as a private dataset? Asking because my submission threw a 'Notebook Exceeded Allowed Compute\nError' when I used one.",
      "votes": null
    },
    {
      "id": "1402077",
      "postDate": "07/27/2021 20:27:14",
      "content": "<p>You can load models from a private data set.  Some models are large and you still have risk of exceeding one of the compute restrictions so you may need to open/close them one at a time.</p>",
      "rawMarkdown": "You can load models from a private data set.  Some models are large and you still have risk of exceeding one of the compute restrictions so you may need to open/close them one at a time.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1382487,
      "author_name": "pcjimmmy",
      "author_url": "",
      "post_date": "07/09/2021 23:20:02",
      "content": "<p>Time limit is for the last submission - for the inference.    You can spend a decade off line training a model.</p>\n<p>And yes - this topic has been asked and answered in almost every competition I have been in for the past years.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1382647,
          "author_name": "jellyz9",
          "author_url": "",
          "post_date": "07/10/2021 05:23:39",
          "content": "<p>I understood.<br>\nThank you very much!</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1382657,
          "author_name": "pcjimmmy",
          "author_url": "",
          "post_date": "07/10/2021 05:41:19",
          "content": "<p>Your doing well so far for your first Kaggle - been here for several years.  Pretty much always a helpful group of people who participate in these competitions.  Made many trips to Japan over the years to visit at Toyota and Honda - never got north to see your part of Japan.  It's a beautiful country.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1402033,
          "author_name": "akshayb7",
          "author_url": "",
          "post_date": "07/27/2021 18:41:14",
          "content": "<p>Hey Jelly, I have a similar concern. Can we upload our models as a private dataset? Asking because my submission threw a 'Notebook Exceeded Allowed Compute<br>\nError' when I used one. </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1402077,
          "author_name": "pcjimmmy",
          "author_url": "",
          "post_date": "07/27/2021 20:27:14",
          "content": "<p>You can load models from a private data set.  Some models are large and you still have risk of exceeding one of the compute restrictions so you may need to open/close them one at a time.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1382445": "Hi,\n\nThis is my first time participating in a code competition, and I'm not sure if I understand the rules correctly, so please let me know.\n\nIn the rules section of the competition, there is a statement that says:\n\n> CPU Notebook <= 6 hours run-time\n> GPU Notebook <= 6 hours run-time\n> Internet access disabled\n> Freely & publicly available external data is allowed, including pre-trained models\n> Submissions must be made using the mlb python module\n\nIs this time limit required only for inference?\n\nIn other words, if I spend a long time (more than 6 hours) training a model locally (or in an environment other than KaggleNotebook), upload the trained model to Kaggle as a PrivateDataset, and then use that model to perform inference in less than 6 hours on KaggleNotebook, does this satisfy the rule?\n\nLet me know if I should ask in a different place.\nI’m sorry if a similar topic has already been posted.\n\nThanks in advance,",
    "1382487": "Time limit is for the last submission - for the inference.    You can spend a decade off line training a model.\n\nAnd yes - this topic has been asked and answered in almost every competition I have been in for the past years.",
    "1382647": "I understood.\nThank you very much!",
    "1382657": "Your doing well so far for your first Kaggle - been here for several years.  Pretty much always a helpful group of people who participate in these competitions.  Made many trips to Japan over the years to visit at Toyota and Honda - never got north to see your part of Japan.  It's a beautiful country.",
    "1402033": "Hey Jelly, I have a similar concern. Can we upload our models as a private dataset? Asking because my submission threw a 'Notebook Exceeded Allowed Compute\nError' when I used one.",
    "1402077": "You can load models from a private data set.  Some models are large and you still have risk of exceeding one of the compute restrictions so you may need to open/close them one at a time."
  },
  "source": "meta"
}