{
  "id": 247500,
  "title": "Loading pre-trained models",
  "url": "/competitions/mlb-player-digital-engagement-forecasting/discussion/247500",
  "author_name": "",
  "post_date": "2021-06-19T23:09:19.327171900Z",
  "votes": 6,
  "comment_count": 2,
  "views": 0,
  "content": "<p>This is my first competition in Kaggle, and I was wondering. Is it possible to load weights from a previous trained neural network, so I can keep training my model bypassing the 6 hours time limit?  Is it legal to do that? <br>\nMy model takes too long to train and I thought of doing this before considering another approach.</p>",
  "messages": [
    {
      "id": "1357717",
      "postDate": "06/19/2021 23:09:19",
      "content": "<p>This is my first competition in Kaggle, and I was wondering. Is it possible to load weights from a previous trained neural network, so I can keep training my model bypassing the 6 hours time limit?  Is it legal to do that? <br>\nMy model takes too long to train and I thought of doing this before considering another approach.</p>",
      "rawMarkdown": "This is my first competition in Kaggle, and I was wondering. Is it possible to load weights from a previous trained neural network, so I can keep training my model bypassing the 6 hours time limit?  Is it legal to do that? \nMy model takes too long to train and I thought of doing this before considering another approach.",
      "votes": null
    },
    {
      "id": "1357726",
      "postDate": "06/19/2021 23:28:38",
      "content": "<p>Yes, you can create a Dataset and upload the weights you've trained locally and then connect your Dataset to your notebook by clicking the \"Add Data\" button on the top right of your notebook.</p>",
      "rawMarkdown": "Yes, you can create a Dataset and upload the weights you've trained locally and then connect your Dataset to your notebook by clicking the \"Add Data\" button on the top right of your notebook.",
      "votes": null
    },
    {
      "id": "1359668",
      "postDate": "06/21/2021 12:34:07",
      "content": "<p>Thanks, I will do that. Sorry for replying late btw, I literally coudn't find the reply button 😅.</p>",
      "rawMarkdown": "Thanks, I will do that. Sorry for replying late btw, I literally coudn't find the reply button 😅.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1357726,
      "author_name": "brandenkmurray",
      "author_url": "",
      "post_date": "06/19/2021 23:28:38",
      "content": "<p>Yes, you can create a Dataset and upload the weights you've trained locally and then connect your Dataset to your notebook by clicking the \"Add Data\" button on the top right of your notebook.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1359668,
          "author_name": "alador",
          "author_url": "",
          "post_date": "06/21/2021 12:34:07",
          "content": "<p>Thanks, I will do that. Sorry for replying late btw, I literally coudn't find the reply button 😅.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1357717": "This is my first competition in Kaggle, and I was wondering. Is it possible to load weights from a previous trained neural network, so I can keep training my model bypassing the 6 hours time limit?  Is it legal to do that? \nMy model takes too long to train and I thought of doing this before considering another approach.",
    "1357726": "Yes, you can create a Dataset and upload the weights you've trained locally and then connect your Dataset to your notebook by clicking the \"Add Data\" button on the top right of your notebook.",
    "1359668": "Thanks, I will do that. Sorry for replying late btw, I literally coudn't find the reply button 😅."
  },
  "source": "meta"
}