{
  "id": 458877,
  "title": "Trying to understand code only submission guideline",
  "url": "/competitions/blood-vessel-segmentation/discussion/458877",
  "author_name": "",
  "post_date": "2023-12-02T04:53:03.244019800Z",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Can I code and train my model anywhere and just do the submission in the notebook. To be clear, pro process flow I am looking at:</p>\n<ol>\n<li>Download train data and train model at my GPUs etc or any place. </li>\n<li>Use the model from above and run inference on the test set on Kaggle notebook submission. </li>\n</ol>\n<p>Is this a valid flow?</p>",
  "messages": [
    {
      "id": "2546064",
      "postDate": "12/02/2023 04:53:03",
      "content": "<p>Can I code and train my model anywhere and just do the submission in the notebook. To be clear, pro process flow I am looking at:</p>\n<ol>\n<li>Download train data and train model at my GPUs etc or any place. </li>\n<li>Use the model from above and run inference on the test set on Kaggle notebook submission. </li>\n</ol>\n<p>Is this a valid flow?</p>",
      "rawMarkdown": "Can I code and train my model anywhere and just do the submission in the notebook. To be clear, pro process flow I am looking at:\n1. Download train data and train model at my GPUs etc or any place. \n2. Use the model from above and run inference on the test set on Kaggle notebook submission. \n\nIs this a valid flow?",
      "votes": null
    },
    {
      "id": "2546514",
      "postDate": "12/02/2023 13:54:10",
      "content": "<p>Yes… it means that for submission you have to run your inference notebook</p>",
      "rawMarkdown": "Yes... it means that for submission you have to run your inference notebook",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2546514,
      "author_name": "anitho2910",
      "author_url": "",
      "post_date": "12/02/2023 13:54:10",
      "content": "<p>Yes… it means that for submission you have to run your inference notebook</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2546064": "Can I code and train my model anywhere and just do the submission in the notebook. To be clear, pro process flow I am looking at:\n1. Download train data and train model at my GPUs etc or any place. \n2. Use the model from above and run inference on the test set on Kaggle notebook submission. \n\nIs this a valid flow?",
    "2546514": "Yes... it means that for submission you have to run your inference notebook"
  },
  "source": "meta"
}