{
  "id": 421215,
  "title": "Is it possible to use own pre-trained network data?",
  "url": "/competitions/google-research-identify-contrails-reduce-global-warming/discussion/421215",
  "author_name": "Maxim Podorov",
  "post_date": "2023-07-04T11:49:15.740000",
  "votes": 2,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I want to participate in this competition, but I can't train my network within the 9 hours limit. Is it legal to train a model before making submission, then post its weights as a public dataset in encrypted format (so that nobody except me would be able to use them) and use this dataset in my competition notebook?</p>",
  "messages": [
    {
      "id": 2329621,
      "postDate": "2023-07-04T11:49:15.740Z",
      "content": "<p>I want to participate in this competition, but I can't train my network within the 9 hours limit. Is it legal to train a model before making submission, then post its weights as a public dataset in encrypted format (so that nobody except me would be able to use them) and use this dataset in my competition notebook?</p>",
      "rawMarkdown": "I want to participate in this competition, but I can't train my network within the 9 hours limit. Is it legal to train a model before making submission, then post its weights as a public dataset in encrypted format (so that nobody except me would be able to use them) and use this dataset in my competition notebook?",
      "votes": 1
    },
    {
      "id": 2330176,
      "postDate": "2023-07-04T19:21:31.217Z",
      "content": "<p>Yes, it is allowed, as far as I could tell it is a common practice. And moreover you could upload the models to a private dataset without need for encryption</p>",
      "rawMarkdown": "Yes, it is allowed, as far as I could tell it is a common practice. And moreover you could upload the models to a private dataset without need for encryption",
      "votes": 2,
      "replies": [
        {
          "id": 2331746,
          "postDate": "2023-07-05T18:43:04.050Z",
          "content": "<p>Thank you.</p>",
          "rawMarkdown": "Thank you."
        }
      ]
    }
  ],
  "comments": [
    {
      "id": 2330176,
      "author_name": "Mikhail Kotyushev",
      "author_url": "",
      "post_date": "2023-07-04T19:21:31.217000",
      "content": "<p>Yes, it is allowed, as far as I could tell it is a common practice. And moreover you could upload the models to a private dataset without need for encryption</p>",
      "votes": 2,
      "replies": [
        {
          "id": 2331746,
          "author_name": "Maxim Podorov",
          "author_url": "",
          "post_date": "2023-07-05T18:43:04.050000",
          "content": "<p>Thank you.</p>",
          "votes": 0,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2329621": "I want to participate in this competition, but I can't train my network within the 9 hours limit. Is it legal to train a model before making submission, then post its weights as a public dataset in encrypted format (so that nobody except me would be able to use them) and use this dataset in my competition notebook?",
    "2330176": "Yes, it is allowed, as far as I could tell it is a common practice. And moreover you could upload the models to a private dataset without need for encryption"
  }
}