{
  "id": 371733,
  "title": "Model training ~ Code requirements ",
  "url": "/competitions/rsna-breast-cancer-detection/discussion/371733",
  "author_name": "Rohit Kumar",
  "post_date": "2022-12-12T01:35:23.769000",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Can we train a model on our local compute and then upload those weights to the notebook on kaggle? Would that satisfy the code requirements? </p>\n<p>From what I have understood, only the code required to create the submission file should be available in the notebook and that should compute within 9 hours! Please correct me if I am wrong! </p>",
  "messages": [
    {
      "id": 2062308,
      "postDate": "2022-12-12T01:35:23.770Z",
      "content": "<p>Can we train a model on our local compute and then upload those weights to the notebook on kaggle? Would that satisfy the code requirements? </p>\n<p>From what I have understood, only the code required to create the submission file should be available in the notebook and that should compute within 9 hours! Please correct me if I am wrong! </p>",
      "rawMarkdown": "Can we train a model on our local compute and then upload those weights to the notebook on kaggle? Would that satisfy the code requirements? \n\nFrom what I have understood, only the code required to create the submission file should be available in the notebook and that should compute within 9 hours! Please correct me if I am wrong! ",
      "votes": 1
    },
    {
      "id": 2062361,
      "postDate": "2022-12-12T03:30:28.407Z",
      "content": "<p>You only just use the model in your infer less than 9h. It's OK. </p>",
      "rawMarkdown": "You only just use the model in your infer less than 9h. It's OK. "
    }
  ],
  "comments": [
    {
      "id": 2062361,
      "author_name": "柯慕灵",
      "author_url": "",
      "post_date": "2022-12-12T03:30:28.407000",
      "content": "<p>You only just use the model in your infer less than 9h. It's OK. </p>",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2062308": "Can we train a model on our local compute and then upload those weights to the notebook on kaggle? Would that satisfy the code requirements? \n\nFrom what I have understood, only the code required to create the submission file should be available in the notebook and that should compute within 9 hours! Please correct me if I am wrong! ",
    "2062361": "You only just use the model in your infer less than 9h. It's OK. "
  }
}