{
  "id": 198407,
  "title": "Why is internet disabled for submission?",
  "url": "/competitions/hubmap-kidney-segmentation/discussion/198407",
  "author_name": "",
  "post_date": "2020-11-21T05:34:42.033440400Z",
  "votes": 5,
  "comment_count": 11,
  "views": 0,
  "content": "<p>I'm using a few external libraries (<code>pyvips efficientnet segmentation-models</code>) that require internet connection. It would be nice if we have an easy way to install those libraries, especially since pyvips requires <code>apt install</code>, as mentioned in <a href=\"https://www.kaggle.com/c/hubmap-kidney-segmentation/discussion/198050\" target=\"_blank\">this thread</a>.</p>",
  "messages": [
    {
      "id": "1085694",
      "postDate": "11/21/2020 05:34:42",
      "content": "<p>I'm using a few external libraries (<code>pyvips efficientnet segmentation-models</code>) that require internet connection. It would be nice if we have an easy way to install those libraries, especially since pyvips requires <code>apt install</code>, as mentioned in <a href=\"https://www.kaggle.com/c/hubmap-kidney-segmentation/discussion/198050\" target=\"_blank\">this thread</a>.</p>",
      "rawMarkdown": "I'm using a few external libraries (`pyvips efficientnet segmentation-models`) that require internet connection. It would be nice if we have an easy way to install those libraries, especially since pyvips requires `apt install`, as mentioned in [this thread](https://www.kaggle.com/c/hubmap-kidney-segmentation/discussion/198050).",
      "votes": null
    },
    {
      "id": "1085719",
      "postDate": "11/21/2020 06:29:06",
      "content": "<p>We host code-only competitions in situations where we believe the private test set should remain hidden (for a number of reasons). We must disable internet access to prevent a variety of potential loopholes that could take advantage of the test set. </p>\n<p>The best option would be to request the libraries be added to our kernel images, which would then no longer require internet access :)</p>",
      "rawMarkdown": "We host code-only competitions in situations where we believe the private test set should remain hidden (for a number of reasons). We must disable internet access to prevent a variety of potential loopholes that could take advantage of the test set. \n\nThe best option would be to request the libraries be added to our kernel images, which would then no longer require internet access :)",
      "votes": null
    },
    {
      "id": "1085729",
      "postDate": "11/21/2020 06:42:29",
      "content": "<p>Awesome! It would be great if pyvips is added</p>",
      "rawMarkdown": "Awesome! It would be great if pyvips is added",
      "votes": null
    },
    {
      "id": "1085783",
      "postDate": "11/21/2020 07:38:19",
      "content": "<p>Another option is to add the source or a .whl of the library you want as a Kaggle dataset. Then you can add the dataset to your kernel and install from there</p>",
      "rawMarkdown": "Another option is to add the source or a .whl of the library you want as a Kaggle dataset. Then you can add the dataset to your kernel and install from there",
      "votes": null
    },
    {
      "id": "1086260",
      "postDate": "11/21/2020 14:21:02",
      "content": "<p>Then, we can't install libraries for this competiton?</p>",
      "rawMarkdown": "Then, we can't install libraries for this competiton?",
      "votes": null
    },
    {
      "id": "1086559",
      "postDate": "11/21/2020 19:22:49",
      "content": "<p>Yeah it works fine for <code>pip</code> only library, but since <code>pyvis</code> requires <code>apt-get</code> some libraries it become cumbersome to figure out how to store those files.</p>",
      "rawMarkdown": "Yeah it works fine for `pip` only library, but since `pyvis` requires `apt-get` some libraries it become cumbersome to figure out how to store those files.",
      "votes": null
    },
    {
      "id": "1086610",
      "postDate": "11/21/2020 20:50:28",
      "content": "<p>Oh you're right, sorry I didn't read your original post clearly enough</p>",
      "rawMarkdown": "Oh you're right, sorry I didn't read your original post clearly enough",
      "votes": null
    },
    {
      "id": "1086661",
      "postDate": "11/21/2020 23:02:04",
      "content": "<p>No worries I wasn't very clear in my post anyway</p>",
      "rawMarkdown": "No worries I wasn't very clear in my post anyway",
      "votes": null
    },
    {
      "id": "1086890",
      "postDate": "11/22/2020 06:43:05",
      "content": "<p>Check out this discussion, if anyone looking to install packages offline with PIP.</p>\n<p><a href=\"https://www.kaggle.com/c/severstal-steel-defect-detection/discussion/109679\" target=\"_blank\">https://www.kaggle.com/c/severstal-steel-defect-detection/discussion/109679</a></p>",
      "rawMarkdown": "Check out this discussion, if anyone looking to install packages offline with PIP.\n\nhttps://www.kaggle.com/c/severstal-steel-defect-detection/discussion/109679",
      "votes": null
    },
    {
      "id": "1088459",
      "postDate": "11/23/2020 16:20:19",
      "content": "<p>I have made a <a href=\"https://www.kaggle.com/vineeth1999/hubmap-eda-pytorch-efficientunet-offline-training\" target=\"_blank\">kernel</a> which includes the updated version of my segmentation models <a href=\"https://www.kaggle.com/vineeth1999/segmentationmodelspytorch\" target=\"_blank\">dataset</a> for pytorch for offline installation of the packages.</p>",
      "rawMarkdown": "I have made a [kernel](https://www.kaggle.com/vineeth1999/hubmap-eda-pytorch-efficientunet-offline-training) which includes the updated version of my segmentation models [dataset](https://www.kaggle.com/vineeth1999/segmentationmodelspytorch) for pytorch for offline installation of the packages.",
      "votes": null
    },
    {
      "id": "1092627",
      "postDate": "11/27/2020 03:47:44",
      "content": "<p>is it possible if we install the library to<code>/kaggle/working</code> and make the script as utility will solve i think</p>",
      "rawMarkdown": "is it possible if we install the library to` /kaggle/working` and make the script as utility will solve i think",
      "votes": null
    },
    {
      "id": "1094932",
      "postDate": "11/29/2020 05:21:34",
      "content": "<p>check this <a href=\"https://www.kaggle.com/c/hubmap-kidney-segmentation/discussion/200010#1094411\" target=\"_blank\">link</a> for how i did it. you can install anylibrary with pip also from github represotory </p>",
      "rawMarkdown": "check this [link](https://www.kaggle.com/c/hubmap-kidney-segmentation/discussion/200010#1094411) for how i did it. you can install anylibrary with pip also from github represotory",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1085719,
      "author_name": "addisonhoward",
      "author_url": "",
      "post_date": "11/21/2020 06:29:06",
      "content": "<p>We host code-only competitions in situations where we believe the private test set should remain hidden (for a number of reasons). We must disable internet access to prevent a variety of potential loopholes that could take advantage of the test set. </p>\n<p>The best option would be to request the libraries be added to our kernel images, which would then no longer require internet access :)</p>",
      "votes": null,
      "replies": [
        {
          "id": 1085729,
          "author_name": "xhlulu",
          "author_url": "",
          "post_date": "11/21/2020 06:42:29",
          "content": "<p>Awesome! It would be great if pyvips is added</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1085783,
          "author_name": "matthewmasters",
          "author_url": "",
          "post_date": "11/21/2020 07:38:19",
          "content": "<p>Another option is to add the source or a .whl of the library you want as a Kaggle dataset. Then you can add the dataset to your kernel and install from there</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1086559,
          "author_name": "xhlulu",
          "author_url": "",
          "post_date": "11/21/2020 19:22:49",
          "content": "<p>Yeah it works fine for <code>pip</code> only library, but since <code>pyvis</code> requires <code>apt-get</code> some libraries it become cumbersome to figure out how to store those files.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1086610,
          "author_name": "matthewmasters",
          "author_url": "",
          "post_date": "11/21/2020 20:50:28",
          "content": "<p>Oh you're right, sorry I didn't read your original post clearly enough</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1086661,
          "author_name": "xhlulu",
          "author_url": "",
          "post_date": "11/21/2020 23:02:04",
          "content": "<p>No worries I wasn't very clear in my post anyway</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1092627,
          "author_name": "morizin",
          "author_url": "",
          "post_date": "11/27/2020 03:47:44",
          "content": "<p>is it possible if we install the library to<code>/kaggle/working</code> and make the script as utility will solve i think</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1094932,
          "author_name": "morizin",
          "author_url": "",
          "post_date": "11/29/2020 05:21:34",
          "content": "<p>check this <a href=\"https://www.kaggle.com/c/hubmap-kidney-segmentation/discussion/200010#1094411\" target=\"_blank\">link</a> for how i did it. you can install anylibrary with pip also from github represotory </p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1086260,
      "author_name": "hiramcho",
      "author_url": "",
      "post_date": "11/21/2020 14:21:02",
      "content": "<p>Then, we can't install libraries for this competiton?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1086890,
      "author_name": "joshi98kishan",
      "author_url": "",
      "post_date": "11/22/2020 06:43:05",
      "content": "<p>Check out this discussion, if anyone looking to install packages offline with PIP.</p>\n<p><a href=\"https://www.kaggle.com/c/severstal-steel-defect-detection/discussion/109679\" target=\"_blank\">https://www.kaggle.com/c/severstal-steel-defect-detection/discussion/109679</a></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1088459,
      "author_name": "vineeth1999",
      "author_url": "",
      "post_date": "11/23/2020 16:20:19",
      "content": "<p>I have made a <a href=\"https://www.kaggle.com/vineeth1999/hubmap-eda-pytorch-efficientunet-offline-training\" target=\"_blank\">kernel</a> which includes the updated version of my segmentation models <a href=\"https://www.kaggle.com/vineeth1999/segmentationmodelspytorch\" target=\"_blank\">dataset</a> for pytorch for offline installation of the packages.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1085694": "I'm using a few external libraries (`pyvips efficientnet segmentation-models`) that require internet connection. It would be nice if we have an easy way to install those libraries, especially since pyvips requires `apt install`, as mentioned in [this thread](https://www.kaggle.com/c/hubmap-kidney-segmentation/discussion/198050).",
    "1085719": "We host code-only competitions in situations where we believe the private test set should remain hidden (for a number of reasons). We must disable internet access to prevent a variety of potential loopholes that could take advantage of the test set. \n\nThe best option would be to request the libraries be added to our kernel images, which would then no longer require internet access :)",
    "1085729": "Awesome! It would be great if pyvips is added",
    "1085783": "Another option is to add the source or a .whl of the library you want as a Kaggle dataset. Then you can add the dataset to your kernel and install from there",
    "1086260": "Then, we can't install libraries for this competiton?",
    "1086559": "Yeah it works fine for `pip` only library, but since `pyvis` requires `apt-get` some libraries it become cumbersome to figure out how to store those files.",
    "1086610": "Oh you're right, sorry I didn't read your original post clearly enough",
    "1086661": "No worries I wasn't very clear in my post anyway",
    "1086890": "Check out this discussion, if anyone looking to install packages offline with PIP.\n\nhttps://www.kaggle.com/c/severstal-steel-defect-detection/discussion/109679",
    "1088459": "I have made a [kernel](https://www.kaggle.com/vineeth1999/hubmap-eda-pytorch-efficientunet-offline-training) which includes the updated version of my segmentation models [dataset](https://www.kaggle.com/vineeth1999/segmentationmodelspytorch) for pytorch for offline installation of the packages.",
    "1092627": "is it possible if we install the library to` /kaggle/working` and make the script as utility will solve i think",
    "1094932": "check this [link](https://www.kaggle.com/c/hubmap-kidney-segmentation/discussion/200010#1094411) for how i did it. you can install anylibrary with pip also from github represotory"
  },
  "source": "meta"
}