{
  "id": 575622,
  "title": "Adding a model as input in a notebook",
  "url": "/competitions/byu-locating-bacterial-flagellar-motors-2025/discussion/575622",
  "author_name": "",
  "post_date": "2025-04-29T21:55:48.131435200Z",
  "votes": null,
  "comment_count": 3,
  "views": 0,
  "content": "<p>I want to continue training a model that has already been trained for another 50 epochs. Is it possible to use this method for the competition file?</p>",
  "messages": [
    {
      "id": "3189865",
      "postDate": "04/29/2025 21:55:48",
      "content": "<p>I want to continue training a model that has already been trained for another 50 epochs. Is it possible to use this method for the competition file?</p>",
      "rawMarkdown": "I want to continue training a model that has already been trained for another 50 epochs. Is it possible to use this method for the competition file?",
      "votes": null
    },
    {
      "id": "3189899",
      "postDate": "04/29/2025 22:49:49",
      "content": "<p>U can just load weights and continue training </p>",
      "rawMarkdown": "U can just load weights and continue training",
      "votes": null
    },
    {
      "id": "3190429",
      "postDate": "04/30/2025 16:55:23",
      "content": "<p>Do not train a model in your submission notebook.  Train your model to completion and then save the weights as a .pt file.  Put the .pt file into a dataset and load it into your notebook.  Training on your submission wastes time and cannot guarantee the best model is submitted (due to slight randomness when training).  I highly recommend training in a different environment than your submission notebook.</p>",
      "rawMarkdown": "Do not train a model in your submission notebook.  Train your model to completion and then save the weights as a .pt file.  Put the .pt file into a dataset and load it into your notebook.  Training on your submission wastes time and cannot guarantee the best model is submitted (due to slight randomness when training).  I highly recommend training in a different environment than your submission notebook.",
      "votes": null
    },
    {
      "id": "3190617",
      "postDate": "04/30/2025 23:42:25",
      "content": "<p>Hi, did you find out how to do this?</p>",
      "rawMarkdown": "Hi, did you find out how to do this?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3189899,
      "author_name": "eikyou",
      "author_url": "",
      "post_date": "04/29/2025 22:49:49",
      "content": "<p>U can just load weights and continue training </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 3190429,
      "author_name": "connorjd",
      "author_url": "",
      "post_date": "04/30/2025 16:55:23",
      "content": "<p>Do not train a model in your submission notebook.  Train your model to completion and then save the weights as a .pt file.  Put the .pt file into a dataset and load it into your notebook.  Training on your submission wastes time and cannot guarantee the best model is submitted (due to slight randomness when training).  I highly recommend training in a different environment than your submission notebook.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 3190617,
      "author_name": "kodevirtuoso",
      "author_url": "",
      "post_date": "04/30/2025 23:42:25",
      "content": "<p>Hi, did you find out how to do this?</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3189865": "I want to continue training a model that has already been trained for another 50 epochs. Is it possible to use this method for the competition file?",
    "3189899": "U can just load weights and continue training",
    "3190429": "Do not train a model in your submission notebook.  Train your model to completion and then save the weights as a .pt file.  Put the .pt file into a dataset and load it into your notebook.  Training on your submission wastes time and cannot guarantee the best model is submitted (due to slight randomness when training).  I highly recommend training in a different environment than your submission notebook.",
    "3190617": "Hi, did you find out how to do this?"
  },
  "source": "meta"
}