{
  "id": 401441,
  "title": "Can I save the trained model and directly use the model to process the test dataset in the code submission?",
  "url": "/competitions/icecube-neutrinos-in-deep-ice/discussion/401441",
  "author_name": "Starp",
  "post_date": "2023-04-13T07:47:09.626000",
  "votes": 0,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Hello, it's my first time trying a \"code competition\".<br>\nI wonder whether it is OK to save my trained model in the workspace and just apply it to the test dataset in the notebook for submission.<br>\nOr the whole process (data process, train, apply) must be included in the submission?</p>",
  "messages": [
    {
      "id": 2220236,
      "postDate": "2023-04-13T08:00:37.150Z",
      "content": "<p>You can do that. You can also train locally and upload your model weights as a dataset and use that for inference. No need to include the training code in your submission. Just the inference code is needed.</p>",
      "rawMarkdown": "You can do that. You can also train locally and upload your model weights as a dataset and use that for inference. No need to include the training code in your submission. Just the inference code is needed.",
      "votes": 1,
      "replies": [
        {
          "id": 2220248,
          "postDate": "2023-04-13T08:09:22.083Z",
          "content": "<p>Thank you very much for the explanation!</p>",
          "rawMarkdown": "Thank you very much for the explanation!",
          "votes": 1
        }
      ]
    },
    {
      "id": 2220217,
      "postDate": "2023-04-13T07:47:09.627Z",
      "content": "<p>Hello, it's my first time trying a \"code competition\".<br>\nI wonder whether it is OK to save my trained model in the workspace and just apply it to the test dataset in the notebook for submission.<br>\nOr the whole process (data process, train, apply) must be included in the submission?</p>",
      "rawMarkdown": "Hello, it's my first time trying a \"code competition\".\nI wonder whether it is OK to save my trained model in the workspace and just apply it to the test dataset in the notebook for submission.\nOr the whole process (data process, train, apply) must be included in the submission?"
    },
    {
      "id": 2221263,
      "postDate": "2023-04-14T05:23:56.190Z",
      "rawMarkdown": "",
      "isDeleted": true
    }
  ],
  "comments": [
    {
      "id": 2220236,
      "author_name": "CroDoc",
      "author_url": "",
      "post_date": "2023-04-13T08:00:37.150000",
      "content": "<p>You can do that. You can also train locally and upload your model weights as a dataset and use that for inference. No need to include the training code in your submission. Just the inference code is needed.</p>",
      "votes": 1,
      "replies": [
        {
          "id": 2220248,
          "author_name": "Starp",
          "author_url": "",
          "post_date": "2023-04-13T08:09:22.083000",
          "content": "<p>Thank you very much for the explanation!</p>",
          "votes": 1,
          "replies": []
        }
      ]
    },
    {
      "id": 2221263,
      "author_name": "",
      "author_url": "",
      "post_date": "2023-04-14T05:23:56.190000",
      "content": "",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2220236": "You can do that. You can also train locally and upload your model weights as a dataset and use that for inference. No need to include the training code in your submission. Just the inference code is needed.",
    "2220217": "Hello, it's my first time trying a \"code competition\".\nI wonder whether it is OK to save my trained model in the workspace and just apply it to the test dataset in the notebook for submission.\nOr the whole process (data process, train, apply) must be included in the submission?",
    "2221263": ""
  }
}