{
  "id": 667669,
  "title": "Help with snntorch or any other package for submit",
  "url": "/competitions/recodai-luc-scientific-image-forgery-detection/discussion/667669",
  "author_name": "",
  "post_date": "2026-01-13T22:51:55.435906700Z",
  "votes": null,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Does anyone know how to install a package and let it stay through (internet off) submit process?\nI am trying to use snntorch. But it is not one of the standard packages. Please help.</p>",
  "messages": [
    {
      "id": "3390758",
      "postDate": "01/13/2026 22:51:55",
      "content": "<p>Does anyone know how to install a package and let it stay through (internet off) submit process?\nI am trying to use snntorch. But it is not one of the standard packages. Please help.</p>",
      "rawMarkdown": "Does anyone know how to install a package and let it stay through (internet off) submit process?\nI am trying to use snntorch. But it is not one of the standard packages. Please help.",
      "votes": null
    },
    {
      "id": "3390765",
      "postDate": "01/13/2026 23:12:49",
      "content": "<p>I am able to solve this by uploading the .whl of the package (and its dependencies) as an external dataset. Then I can install them offline.</p>",
      "rawMarkdown": "I am able to solve this by uploading the .whl of the package (and its dependencies) as an external dataset. Then I can install them offline.",
      "votes": null
    },
    {
      "id": "3390773",
      "postDate": "01/13/2026 23:31:33",
      "content": "<p>Found the solution. \nEssentially, you have to pre download and bundle the package (not installing) to a local directory and upload to kaggle as a dataset.  Then use that newly created dataset path to install the package in internet off mode.</p>\n<p>Below is the step by step process:</p>\n<ol>\n<li><p>Create a Temporary Notebook with Internet ON\nOpen a new Kaggle notebook.\nEnable Internet in the settings.\nInstall the package you need and save it as a .whl or .tar.gz file.</p>\n<pre><code>!pip download mypackage==1.2.3 -d ./packages<pre><code>This will download the package and its dependencies into the ./packages folder.\n</code></pre></code></pre></li>\n<li><p>Save the Package Files to Kaggle Dataset\nGo to the Data tab in Kaggle.\nClick Create New Dataset.\nUpload the ./packages folder you downloaded.\nGive it a name like mypackage-offline.</p></li>\n<li><p>Use the Dataset in Your Internet-OFF Notebook\nIn your competition notebook (Internet OFF), go to Add Data and attach your dataset mypackage-offline.\nInstall from the local path instead of PyPI:</p>\n<pre><code>!pip install --no-index --find-links=/kaggle/input/mypackage-offline mypackage<pre><code>--no-index ensures it doesn’t try to connect to PyPI.\n--find-links points to the folder containing your .whl files.\n</code></pre></code></pre></li>\n<li><p>Verify Before Submission</p></li>\n</ol>",
      "rawMarkdown": "Found the solution. \nEssentially, you have to pre download and bundle the package (not installing) to a local directory and upload to kaggle as a dataset.  Then use that newly created dataset path to install the package in internet off mode.\n\n\nBelow is the step by step process:\n1. Create a Temporary Notebook with Internet ON\n\tOpen a new Kaggle notebook.\n\tEnable Internet in the settings.\n\tInstall the package you need and save it as a .whl or .tar.gz file.\n\t\n\t\t!pip download mypackage==1.2.3 -d ./packages\n\t\t\t\n\t\t\tThis will download the package and its dependencies into the ./packages folder.\n\n2. Save the Package Files to Kaggle Dataset\n\tGo to the Data tab in Kaggle.\n\tClick Create New Dataset.\n\tUpload the ./packages folder you downloaded.\n\tGive it a name like mypackage-offline.\n\n3. Use the Dataset in Your Internet-OFF Notebook\n\tIn your competition notebook (Internet OFF), go to Add Data and attach your dataset mypackage-offline.\n\tInstall from the local path instead of PyPI:\n\n\t\t!pip install --no-index --find-links=/kaggle/input/mypackage-offline mypackage\n\t\t\t\n\t\t\t--no-index ensures it doesn’t try to connect to PyPI.\n\t\t\t--find-links points to the folder containing your .whl files.\n\n4. Verify Before Submission",
      "votes": null
    },
    {
      "id": "3391438",
      "postDate": "01/15/2026 00:09:47",
      "content": "<p>Yes. Thank you.</p>",
      "rawMarkdown": "Yes. Thank you.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3390765,
      "author_name": "returnofsputnik",
      "author_url": "",
      "post_date": "01/13/2026 23:12:49",
      "content": "<p>I am able to solve this by uploading the .whl of the package (and its dependencies) as an external dataset. Then I can install them offline.</p>",
      "votes": null,
      "replies": [
        {
          "id": 3391438,
          "author_name": "venkatagoli",
          "author_url": "",
          "post_date": "01/15/2026 00:09:47",
          "content": "<p>Yes. Thank you.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 3390773,
      "author_name": "venkatagoli",
      "author_url": "",
      "post_date": "01/13/2026 23:31:33",
      "content": "<p>Found the solution. \nEssentially, you have to pre download and bundle the package (not installing) to a local directory and upload to kaggle as a dataset.  Then use that newly created dataset path to install the package in internet off mode.</p>\n<p>Below is the step by step process:</p>\n<ol>\n<li><p>Create a Temporary Notebook with Internet ON\nOpen a new Kaggle notebook.\nEnable Internet in the settings.\nInstall the package you need and save it as a .whl or .tar.gz file.</p>\n<pre><code>!pip download mypackage==1.2.3 -d ./packages<pre><code>This will download the package and its dependencies into the ./packages folder.\n</code></pre></code></pre></li>\n<li><p>Save the Package Files to Kaggle Dataset\nGo to the Data tab in Kaggle.\nClick Create New Dataset.\nUpload the ./packages folder you downloaded.\nGive it a name like mypackage-offline.</p></li>\n<li><p>Use the Dataset in Your Internet-OFF Notebook\nIn your competition notebook (Internet OFF), go to Add Data and attach your dataset mypackage-offline.\nInstall from the local path instead of PyPI:</p>\n<pre><code>!pip install --no-index --find-links=/kaggle/input/mypackage-offline mypackage<pre><code>--no-index ensures it doesn’t try to connect to PyPI.\n--find-links points to the folder containing your .whl files.\n</code></pre></code></pre></li>\n<li><p>Verify Before Submission</p></li>\n</ol>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3390758": "Does anyone know how to install a package and let it stay through (internet off) submit process?\nI am trying to use snntorch. But it is not one of the standard packages. Please help.",
    "3390765": "I am able to solve this by uploading the .whl of the package (and its dependencies) as an external dataset. Then I can install them offline.",
    "3390773": "Found the solution. \nEssentially, you have to pre download and bundle the package (not installing) to a local directory and upload to kaggle as a dataset.  Then use that newly created dataset path to install the package in internet off mode.\n\n\nBelow is the step by step process:\n1. Create a Temporary Notebook with Internet ON\n\tOpen a new Kaggle notebook.\n\tEnable Internet in the settings.\n\tInstall the package you need and save it as a .whl or .tar.gz file.\n\t\n\t\t!pip download mypackage==1.2.3 -d ./packages\n\t\t\t\n\t\t\tThis will download the package and its dependencies into the ./packages folder.\n\n2. Save the Package Files to Kaggle Dataset\n\tGo to the Data tab in Kaggle.\n\tClick Create New Dataset.\n\tUpload the ./packages folder you downloaded.\n\tGive it a name like mypackage-offline.\n\n3. Use the Dataset in Your Internet-OFF Notebook\n\tIn your competition notebook (Internet OFF), go to Add Data and attach your dataset mypackage-offline.\n\tInstall from the local path instead of PyPI:\n\n\t\t!pip install --no-index --find-links=/kaggle/input/mypackage-offline mypackage\n\t\t\t\n\t\t\t--no-index ensures it doesn’t try to connect to PyPI.\n\t\t\t--find-links points to the folder containing your .whl files.\n\n4. Verify Before Submission",
    "3391438": "Yes. Thank you."
  },
  "source": "meta"
}