{
  "id": 582482,
  "title": "Packages for Submission",
  "url": "/competitions/image-matching-challenge-2025/discussion/582482",
  "author_name": "",
  "post_date": "2025-05-31T15:16:04.445427200Z",
  "votes": null,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Hi everyone,</p>\n<p>I'm preparing a Kaggle notebook for an offline submission. When my notebook has an internet connection, I can install all the necessary Python libraries, and their files (like wheels) are downloaded to /kaggle/working.</p>\n<p>I understand that for the actual offline submission, these libraries need to be accessible from /kaggle/input. What's the recommended method to take the libraries I've downloaded to /kaggle/working (while online) and package them into a Kaggle dataset? The goal is to then add this dataset as an input to my submission notebook, ensuring all dependencies are available when it runs offline without internet access.</p>\n<p>Are there any easy or standard ways to manage this process effectively to avoid issues during submission? Thanks!</p>",
  "messages": [
    {
      "id": "3214455",
      "postDate": "05/31/2025 15:16:04",
      "content": "<p>Hi everyone,</p>\n<p>I'm preparing a Kaggle notebook for an offline submission. When my notebook has an internet connection, I can install all the necessary Python libraries, and their files (like wheels) are downloaded to /kaggle/working.</p>\n<p>I understand that for the actual offline submission, these libraries need to be accessible from /kaggle/input. What's the recommended method to take the libraries I've downloaded to /kaggle/working (while online) and package them into a Kaggle dataset? The goal is to then add this dataset as an input to my submission notebook, ensuring all dependencies are available when it runs offline without internet access.</p>\n<p>Are there any easy or standard ways to manage this process effectively to avoid issues during submission? Thanks!</p>",
      "rawMarkdown": "Hi everyone,\n\nI'm preparing a Kaggle notebook for an offline submission. When my notebook has an internet connection, I can install all the necessary Python libraries, and their files (like wheels) are downloaded to /kaggle/working.\n\nI understand that for the actual offline submission, these libraries need to be accessible from /kaggle/input. What's the recommended method to take the libraries I've downloaded to /kaggle/working (while online) and package them into a Kaggle dataset? The goal is to then add this dataset as an input to my submission notebook, ensuring all dependencies are available when it runs offline without internet access.\n\nAre there any easy or standard ways to manage this process effectively to avoid issues during submission? Thanks!",
      "votes": null
    },
    {
      "id": "3214460",
      "postDate": "05/31/2025 15:22:33",
      "content": "<p>Umm, don't put time in this, although some people still prefer the way you are trying to use.</p>\n<p>go to 'Add ons' of your notebook there select install dependencies, there you can write any commands that you want to run before your notebook start running to build the environment.</p>",
      "rawMarkdown": "Umm, don't put time in this, although some people still prefer the way you are trying to use.\n\ngo to 'Add ons' of your notebook there select install dependencies, there you can write any commands that you want to run before your notebook start running to build the environment.",
      "votes": null
    },
    {
      "id": "3214539",
      "postDate": "05/31/2025 18:21:38",
      "content": "<p>Oh, I haven't known that feature until seeing your answer. Thank you very much ^^ I have questions about it.</p>\n<ol>\n<li><p>When I tried \"add-ons\" without internet connection, run button was not active and \"To run the pip installations,please be sure that internet connection is turned on.\" was written. If I understood correctly, it will not like that when I submit the notebook. Am I right?</p></li>\n<li><p>Normally, I had to restart the session when I install the dependencies, will it be a problem or not?</p></li>\n</ol>\n<p>Thank you very much ^^</p>",
      "rawMarkdown": "Oh, I haven't known that feature until seeing your answer. Thank you very much ^^ I have questions about it.\n\n1. When I tried \"add-ons\" without internet connection, run button was not active and \"To run the pip installations,please be sure that internet connection is turned on.\" was written. If I understood correctly, it will not like that when I submit the notebook. Am I right?\n\n2. Normally, I had to restart the session when I install the dependencies, will it be a problem or not?\n\nThank you very much ^^",
      "votes": null
    },
    {
      "id": "3214564",
      "postDate": "05/31/2025 19:34:36",
      "content": "<p>On version saving, it will have execute those commands, even if internet access is disable in your notebook.<br>\nYou need internet access to run that in interactive session only.</p>\n<p>one tip, don't know bug is fixed or not.<br>\nwhen installing libs in interactive session do it usually, internet enable and running code cell, while saving for submission, do what i described before.</p>",
      "rawMarkdown": "On version saving, it will have execute those commands, even if internet access is disable in your notebook.\nYou need internet access to run that in interactive session only.\n\none tip, don't know bug is fixed or not.\nwhen installing libs in interactive session do it usually, internet enable and running code cell, while saving for submission, do what i described before.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3214460,
      "author_name": "arpit1bansal",
      "author_url": "",
      "post_date": "05/31/2025 15:22:33",
      "content": "<p>Umm, don't put time in this, although some people still prefer the way you are trying to use.</p>\n<p>go to 'Add ons' of your notebook there select install dependencies, there you can write any commands that you want to run before your notebook start running to build the environment.</p>",
      "votes": null,
      "replies": [
        {
          "id": 3214539,
          "author_name": "littlestronomer",
          "author_url": "",
          "post_date": "05/31/2025 18:21:38",
          "content": "<p>Oh, I haven't known that feature until seeing your answer. Thank you very much ^^ I have questions about it.</p>\n<ol>\n<li><p>When I tried \"add-ons\" without internet connection, run button was not active and \"To run the pip installations,please be sure that internet connection is turned on.\" was written. If I understood correctly, it will not like that when I submit the notebook. Am I right?</p></li>\n<li><p>Normally, I had to restart the session when I install the dependencies, will it be a problem or not?</p></li>\n</ol>\n<p>Thank you very much ^^</p>",
          "votes": null,
          "replies": [
            {
              "id": 3214564,
              "author_name": "arpit1bansal",
              "author_url": "",
              "post_date": "05/31/2025 19:34:36",
              "content": "<p>On version saving, it will have execute those commands, even if internet access is disable in your notebook.<br>\nYou need internet access to run that in interactive session only.</p>\n<p>one tip, don't know bug is fixed or not.<br>\nwhen installing libs in interactive session do it usually, internet enable and running code cell, while saving for submission, do what i described before.</p>",
              "votes": null,
              "replies": []
            }
          ]
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "3214455": "Hi everyone,\n\nI'm preparing a Kaggle notebook for an offline submission. When my notebook has an internet connection, I can install all the necessary Python libraries, and their files (like wheels) are downloaded to /kaggle/working.\n\nI understand that for the actual offline submission, these libraries need to be accessible from /kaggle/input. What's the recommended method to take the libraries I've downloaded to /kaggle/working (while online) and package them into a Kaggle dataset? The goal is to then add this dataset as an input to my submission notebook, ensuring all dependencies are available when it runs offline without internet access.\n\nAre there any easy or standard ways to manage this process effectively to avoid issues during submission? Thanks!",
    "3214460": "Umm, don't put time in this, although some people still prefer the way you are trying to use.\n\ngo to 'Add ons' of your notebook there select install dependencies, there you can write any commands that you want to run before your notebook start running to build the environment.",
    "3214539": "Oh, I haven't known that feature until seeing your answer. Thank you very much ^^ I have questions about it.\n\n1. When I tried \"add-ons\" without internet connection, run button was not active and \"To run the pip installations,please be sure that internet connection is turned on.\" was written. If I understood correctly, it will not like that when I submit the notebook. Am I right?\n\n2. Normally, I had to restart the session when I install the dependencies, will it be a problem or not?\n\nThank you very much ^^",
    "3214564": "On version saving, it will have execute those commands, even if internet access is disable in your notebook.\nYou need internet access to run that in interactive session only.\n\none tip, don't know bug is fixed or not.\nwhen installing libs in interactive session do it usually, internet enable and running code cell, while saving for submission, do what i described before."
  },
  "source": "meta"
}