{
  "id": 178103,
  "title": "Few questions on training and loading model",
  "url": "/competitions/landmark-recognition-2020/discussion/178103",
  "author_name": "",
  "post_date": "2020-08-28T16:11:47.059589500Z",
  "votes": 2,
  "comment_count": 1,
  "views": 0,
  "content": "<ol>\n<li><p>If I train a model in one script and save it (s1) and load it onto another script (s2) for predictions, is it considered as external data, and do I need to disclose this?</p></li>\n<li><p>When I save a model and make a commit I can see it on the output section. But if i make other commits I am unable to see it anymore. Is there anyway i can load a model which was saved in a previous commit?<br>\n     - this makes training models a bit hard as the kernel times out or after commiting I'm unable to load it again to make updates on the next edit. Can you provide some resource or documentation to help with this problem?</p></li>\n</ol>\n<p><a href=\"https://www.kaggle.com/maggiemd\" target=\"_blank\">@maggiemd</a>  <a href=\"https://www.kaggle.com/camaskew\" target=\"_blank\">@camaskew</a>  <a href=\"https://www.kaggle.com/tobwey\" target=\"_blank\">@tobwey</a> <a href=\"https://www.kaggle.com/andrefaraujo\" target=\"_blank\">@andrefaraujo</a> <a href=\"https://www.kaggle.com/andrefaraujo2\" target=\"_blank\">@andrefaraujo2</a> </p>",
  "messages": [
    {
      "id": "989207",
      "postDate": "08/28/2020 16:11:47",
      "content": "<ol>\n<li><p>If I train a model in one script and save it (s1) and load it onto another script (s2) for predictions, is it considered as external data, and do I need to disclose this?</p></li>\n<li><p>When I save a model and make a commit I can see it on the output section. But if i make other commits I am unable to see it anymore. Is there anyway i can load a model which was saved in a previous commit?<br>\n     - this makes training models a bit hard as the kernel times out or after commiting I'm unable to load it again to make updates on the next edit. Can you provide some resource or documentation to help with this problem?</p></li>\n</ol>\n<p><a href=\"https://www.kaggle.com/maggiemd\" target=\"_blank\">@maggiemd</a>  <a href=\"https://www.kaggle.com/camaskew\" target=\"_blank\">@camaskew</a>  <a href=\"https://www.kaggle.com/tobwey\" target=\"_blank\">@tobwey</a> <a href=\"https://www.kaggle.com/andrefaraujo\" target=\"_blank\">@andrefaraujo</a> <a href=\"https://www.kaggle.com/andrefaraujo2\" target=\"_blank\">@andrefaraujo2</a> </p>",
      "rawMarkdown": "1. If I train a model in one script and save it (s1) and load it onto another script (s2) for predictions, is it considered as external data, and do I need to disclose this?\n\n2. When I save a model and make a commit I can see it on the output section. But if i make other commits I am unable to see it anymore. Is there anyway i can load a model which was saved in a previous commit?\n         - this makes training models a bit hard as the kernel times out or after commiting I'm unable to load it again to make updates on the next edit. Can you provide some resource or documentation to help with this problem?\n\n@maggiemd  @camaskew  @tobwey @andrefaraujo @andrefaraujo2",
      "votes": null
    },
    {
      "id": "1019659",
      "postDate": "09/20/2020 15:22:21",
      "content": "<ol>\n<li>You don't need to disclose anything as you will be importing ur models on ur final notebook.</li>\n<li>What you save goes to the output folder and the notebook doesn't allow you to read that. So you need to import this notebook into another notebook or you can try saving the models in one of the input folders incase it allows.</li>\n</ol>",
      "rawMarkdown": "1. You don't need to disclose anything as you will be importing ur models on ur final notebook.\n2. What you save goes to the output folder and the notebook doesn't allow you to read that. So you need to import this notebook into another notebook or you can try saving the models in one of the input folders incase it allows.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1019659,
      "author_name": "vikrant06",
      "author_url": "",
      "post_date": "09/20/2020 15:22:21",
      "content": "<ol>\n<li>You don't need to disclose anything as you will be importing ur models on ur final notebook.</li>\n<li>What you save goes to the output folder and the notebook doesn't allow you to read that. So you need to import this notebook into another notebook or you can try saving the models in one of the input folders incase it allows.</li>\n</ol>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "989207": "1. If I train a model in one script and save it (s1) and load it onto another script (s2) for predictions, is it considered as external data, and do I need to disclose this?\n\n2. When I save a model and make a commit I can see it on the output section. But if i make other commits I am unable to see it anymore. Is there anyway i can load a model which was saved in a previous commit?\n         - this makes training models a bit hard as the kernel times out or after commiting I'm unable to load it again to make updates on the next edit. Can you provide some resource or documentation to help with this problem?\n\n@maggiemd  @camaskew  @tobwey @andrefaraujo @andrefaraujo2",
    "1019659": "1. You don't need to disclose anything as you will be importing ur models on ur final notebook.\n2. What you save goes to the output folder and the notebook doesn't allow you to read that. So you need to import this notebook into another notebook or you can try saving the models in one of the input folders incase it allows."
  },
  "source": "meta"
}