{
  "id": 458324,
  "title": "Rules regarding code requirements (notebooks)",
  "url": "/competitions/UBC-OCEAN/discussion/458324",
  "author_name": "Nico Varas",
  "post_date": "2023-11-29T09:52:43.376000",
  "votes": 2,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Hi! <br>\nSorry if this has been answered already. I wanted to know if it is \"legal\" in the competition to deliver more than one notebook at the end. For example, having a notebook to train and another notebook to infer. At the end of the competition, will I have to merge them into one? or deliver both as my final submission? <br>\nThanks!</p>",
  "messages": [
    {
      "id": 2542853,
      "postDate": "2023-11-29T15:04:24.790Z",
      "content": "<p>Most people train models somewhere other than Kaggle to avoid memory and time limitations and then just have a notebook for inference to make final submissions.  I personally maintain a GitHub repo of all my code necessary for training and then my notebook is simply me running inference on an already trained model that I have uploaded to Kaggle as a dataset.</p>",
      "rawMarkdown": "Most people train models somewhere other than Kaggle to avoid memory and time limitations and then just have a notebook for inference to make final submissions.  I personally maintain a GitHub repo of all my code necessary for training and then my notebook is simply me running inference on an already trained model that I have uploaded to Kaggle as a dataset.",
      "votes": 4,
      "replies": [
        {
          "id": 2545319,
          "postDate": "2023-12-01T12:37:26.987Z",
          "content": "<p>Thanks a lot! Follow up question: do you recommend any platform to train without the restrictions? </p>",
          "rawMarkdown": "Thanks a lot! Follow up question: do you recommend any platform to train without the restrictions? ",
          "replies": [
            {
              "id": 2551342,
              "postDate": "2023-12-06T16:38:52.683Z",
              "content": "<p>Hello, May be you can try Gradient or Collab</p>",
              "rawMarkdown": "Hello, May be you can try Gradient or Collab",
              "votes": 1
            }
          ]
        }
      ]
    },
    {
      "id": 2542487,
      "postDate": "2023-11-29T09:52:43.377Z",
      "content": "<p>Hi! <br>\nSorry if this has been answered already. I wanted to know if it is \"legal\" in the competition to deliver more than one notebook at the end. For example, having a notebook to train and another notebook to infer. At the end of the competition, will I have to merge them into one? or deliver both as my final submission? <br>\nThanks!</p>",
      "rawMarkdown": "Hi! \nSorry if this has been answered already. I wanted to know if it is \"legal\" in the competition to deliver more than one notebook at the end. For example, having a notebook to train and another notebook to infer. At the end of the competition, will I have to merge them into one? or deliver both as my final submission? \nThanks!",
      "votes": 2
    }
  ],
  "comments": [
    {
      "id": 2542853,
      "author_name": "Connor",
      "author_url": "",
      "post_date": "2023-11-29T15:04:24.790000",
      "content": "<p>Most people train models somewhere other than Kaggle to avoid memory and time limitations and then just have a notebook for inference to make final submissions.  I personally maintain a GitHub repo of all my code necessary for training and then my notebook is simply me running inference on an already trained model that I have uploaded to Kaggle as a dataset.</p>",
      "votes": 4,
      "replies": [
        {
          "id": 2545319,
          "author_name": "Nico Varas",
          "author_url": "",
          "post_date": "2023-12-01T12:37:26.987000",
          "content": "<p>Thanks a lot! Follow up question: do you recommend any platform to train without the restrictions? </p>",
          "votes": 0,
          "replies": [
            {
              "id": 2551342,
              "author_name": "SRIPADAM SUJITH SAI",
              "author_url": "",
              "post_date": "2023-12-06T16:38:52.683000",
              "content": "<p>Hello, May be you can try Gradient or Collab</p>",
              "votes": 1,
              "replies": []
            }
          ]
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2542853": "Most people train models somewhere other than Kaggle to avoid memory and time limitations and then just have a notebook for inference to make final submissions.  I personally maintain a GitHub repo of all my code necessary for training and then my notebook is simply me running inference on an already trained model that I have uploaded to Kaggle as a dataset.",
    "2542487": "Hi! \nSorry if this has been answered already. I wanted to know if it is \"legal\" in the competition to deliver more than one notebook at the end. For example, having a notebook to train and another notebook to infer. At the end of the competition, will I have to merge them into one? or deliver both as my final submission? \nThanks!"
  }
}