{
  "id": 420551,
  "title": "Offline training allowed?",
  "url": "/competitions/hubmap-hacking-the-human-vasculature/discussion/420551",
  "author_name": "",
  "post_date": "2023-07-01T10:30:41.390164100Z",
  "votes": null,
  "comment_count": 5,
  "views": 0,
  "content": "<p>Hey,</p>\n<p>Is it allowed to train with your own data (a.k.a offline training) and only run inference in the notebook with the trained model?</p>",
  "messages": [
    {
      "id": "2325425",
      "postDate": "07/01/2023 10:30:41",
      "content": "<p>Hey,</p>\n<p>Is it allowed to train with your own data (a.k.a offline training) and only run inference in the notebook with the trained model?</p>",
      "rawMarkdown": "Hey,\n\nIs it allowed to train with your own data (a.k.a offline training) and only run inference in the notebook with the trained model?",
      "votes": null
    },
    {
      "id": "2325499",
      "postDate": "07/01/2023 12:14:23",
      "content": "<p>Yes from Code Requirements.</p>\n<p><strong>&gt; Freely &amp; publicly available external data is allowed, including pre-trained models</strong></p>",
      "rawMarkdown": "Yes from Code Requirements.\n\n**> Freely & publicly available external data is allowed, including pre-trained models**",
      "votes": null
    },
    {
      "id": "2325511",
      "postDate": "07/01/2023 12:24:35",
      "content": "<p>But does the \"Freely &amp; publicly available\" also refer to the pre-trained models or only to the external data?</p>",
      "rawMarkdown": "But does the \"Freely & publicly available\" also refer to the pre-trained models or only to the external data?",
      "votes": null
    },
    {
      "id": "2325673",
      "postDate": "07/01/2023 14:15:33",
      "content": "<p>It's \"including pre-trained models\" so I assume it's allowed. However, Kaggle may ask you to submit your code, so it is better to read the \"Rules\".</p>",
      "rawMarkdown": "It's \"including pre-trained models\" so I assume it's allowed. However, Kaggle may ask you to submit your code, so it is better to read the \"Rules\".",
      "votes": null
    },
    {
      "id": "2325690",
      "postDate": "07/01/2023 14:37:11",
      "content": "<p>You can train your model on any device, in your submission notebook import your model's checkpoint from <code>dataset</code> and perform Inference only</p>",
      "rawMarkdown": "You can train your model on any device, in your submission notebook import your model's checkpoint from `dataset` and perform Inference only",
      "votes": null
    },
    {
      "id": "2329214",
      "postDate": "07/04/2023 06:44:10",
      "content": "<p>\"Freely &amp; publicly available\" refers to pre trained models and publicly available external data (such as data from previous competitions), for which you cannot perform any manual annotation</p>",
      "rawMarkdown": "\"Freely & publicly available\" refers to pre trained models and publicly available external data (such as data from previous competitions), for which you cannot perform any manual annotation",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2325499,
      "author_name": "mitsuyasuhoshino",
      "author_url": "",
      "post_date": "07/01/2023 12:14:23",
      "content": "<p>Yes from Code Requirements.</p>\n<p><strong>&gt; Freely &amp; publicly available external data is allowed, including pre-trained models</strong></p>",
      "votes": null,
      "replies": [
        {
          "id": 2325511,
          "author_name": "cookiereg",
          "author_url": "",
          "post_date": "07/01/2023 12:24:35",
          "content": "<p>But does the \"Freely &amp; publicly available\" also refer to the pre-trained models or only to the external data?</p>",
          "votes": null,
          "replies": [
            {
              "id": 2325673,
              "author_name": "mitsuyasuhoshino",
              "author_url": "",
              "post_date": "07/01/2023 14:15:33",
              "content": "<p>It's \"including pre-trained models\" so I assume it's allowed. However, Kaggle may ask you to submit your code, so it is better to read the \"Rules\".</p>",
              "votes": null,
              "replies": []
            },
            {
              "id": 2329214,
              "author_name": "fanshutou",
              "author_url": "",
              "post_date": "07/04/2023 06:44:10",
              "content": "<p>\"Freely &amp; publicly available\" refers to pre trained models and publicly available external data (such as data from previous competitions), for which you cannot perform any manual annotation</p>",
              "votes": null,
              "replies": []
            }
          ]
        }
      ]
    },
    {
      "id": 2325690,
      "author_name": "ptran1203",
      "author_url": "",
      "post_date": "07/01/2023 14:37:11",
      "content": "<p>You can train your model on any device, in your submission notebook import your model's checkpoint from <code>dataset</code> and perform Inference only</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2325425": "Hey,\n\nIs it allowed to train with your own data (a.k.a offline training) and only run inference in the notebook with the trained model?",
    "2325499": "Yes from Code Requirements.\n\n**> Freely & publicly available external data is allowed, including pre-trained models**",
    "2325511": "But does the \"Freely & publicly available\" also refer to the pre-trained models or only to the external data?",
    "2325673": "It's \"including pre-trained models\" so I assume it's allowed. However, Kaggle may ask you to submit your code, so it is better to read the \"Rules\".",
    "2325690": "You can train your model on any device, in your submission notebook import your model's checkpoint from `dataset` and perform Inference only",
    "2329214": "\"Freely & publicly available\" refers to pre trained models and publicly available external data (such as data from previous competitions), for which you cannot perform any manual annotation"
  },
  "source": "meta"
}