{
  "id": 237584,
  "title": "Confusion about the rules",
  "url": "/competitions/plant-pathology-2021-fgvc8/discussion/237584",
  "author_name": "",
  "post_date": "2021-05-09T13:05:59.102250500Z",
  "votes": null,
  "comment_count": 4,
  "views": 0,
  "content": "<p>This is my first Kaggle competition, so please bear with me. I made some nice notebook involving lots of pre-processing that I apply to the images, but when I copy it into a Kaggle notebook I realized that some packages are missing. To install them, I must switch on the internet. Does that mean my notebook violates the rules? Am I only allowed to use the packages that are installed by default?</p>",
  "messages": [
    {
      "id": "1299081",
      "postDate": "05/09/2021 13:05:59",
      "content": "<p>This is my first Kaggle competition, so please bear with me. I made some nice notebook involving lots of pre-processing that I apply to the images, but when I copy it into a Kaggle notebook I realized that some packages are missing. To install them, I must switch on the internet. Does that mean my notebook violates the rules? Am I only allowed to use the packages that are installed by default?</p>",
      "rawMarkdown": "This is my first Kaggle competition, so please bear with me. I made some nice notebook involving lots of pre-processing that I apply to the images, but when I copy it into a Kaggle notebook I realized that some packages are missing. To install them, I must switch on the internet. Does that mean my notebook violates the rules? Am I only allowed to use the packages that are installed by default?",
      "votes": null
    },
    {
      "id": "1299083",
      "postDate": "05/09/2021 13:07:25",
      "content": "<p>No, you can use external packages. But for inference notebook you'll have to include them as Kaggle datasets for installation, since you can't submit your notebook with internet being turned on.</p>",
      "rawMarkdown": "No, you can use external packages. But for inference notebook you'll have to include them as Kaggle datasets for installation, since you can't submit your notebook with internet being turned on.",
      "votes": null
    },
    {
      "id": "1299618",
      "postDate": "05/09/2021 21:37:54",
      "content": "<p>Thanks, I hadn't thought of that. I already found a thread explaining how to do that.</p>",
      "rawMarkdown": "Thanks, I hadn't thought of that. I already found a thread explaining how to do that.",
      "votes": null
    },
    {
      "id": "1306456",
      "postDate": "05/13/2021 19:08:58",
      "content": "<p>share link please</p>",
      "rawMarkdown": "share link please",
      "votes": null
    },
    {
      "id": "1306504",
      "postDate": "05/13/2021 20:25:24",
      "content": "<p><a href=\"https://www.kaggle.com/samuelepino/pip-installing-packages-with-no-internet\" target=\"_blank\">Try this discussion</a></p>\n<p>What I did to successfully (well apart from my score) submit notebooks with external packages:</p>\n<p>1) Open some Kaggle notebook with internet on<br>\n2) run <code>!pip download [insert package name] -d .</code><br>\n3) it will download the package and all its requirements<br>\n4) download all this onto your computer, put it in a zip file<br>\n5) upload it again as Kaggle data into your submission notebook<br>\n6) load it using <code>!pip install [insert package name] --no-index --find-links=file:///kaggle/input/[package folder]</code></p>\n<p>There is probably a smarter way to do it but you only have to do it once per package.</p>",
      "rawMarkdown": "[Try this discussion](https://www.kaggle.com/samuelepino/pip-installing-packages-with-no-internet)\n\nWhat I did to successfully (well apart from my score) submit notebooks with external packages:\n\n1) Open some Kaggle notebook with internet on\n2) run `!pip download [insert package name] -d .`\n3) it will download the package and all its requirements\n4) download all this onto your computer, put it in a zip file\n5) upload it again as Kaggle data into your submission notebook\n6) load it using `!pip install [insert package name] --no-index --find-links=file:///kaggle/input/[package folder]`\n\nThere is probably a smarter way to do it but you only have to do it once per package.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1299083,
      "author_name": "atamazian",
      "author_url": "",
      "post_date": "05/09/2021 13:07:25",
      "content": "<p>No, you can use external packages. But for inference notebook you'll have to include them as Kaggle datasets for installation, since you can't submit your notebook with internet being turned on.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1299618,
      "author_name": "ulrichebling",
      "author_url": "",
      "post_date": "05/09/2021 21:37:54",
      "content": "<p>Thanks, I hadn't thought of that. I already found a thread explaining how to do that.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1306456,
          "author_name": "chtalhaanwar",
          "author_url": "",
          "post_date": "05/13/2021 19:08:58",
          "content": "<p>share link please</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1306504,
          "author_name": "ulrichebling",
          "author_url": "",
          "post_date": "05/13/2021 20:25:24",
          "content": "<p><a href=\"https://www.kaggle.com/samuelepino/pip-installing-packages-with-no-internet\" target=\"_blank\">Try this discussion</a></p>\n<p>What I did to successfully (well apart from my score) submit notebooks with external packages:</p>\n<p>1) Open some Kaggle notebook with internet on<br>\n2) run <code>!pip download [insert package name] -d .</code><br>\n3) it will download the package and all its requirements<br>\n4) download all this onto your computer, put it in a zip file<br>\n5) upload it again as Kaggle data into your submission notebook<br>\n6) load it using <code>!pip install [insert package name] --no-index --find-links=file:///kaggle/input/[package folder]</code></p>\n<p>There is probably a smarter way to do it but you only have to do it once per package.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1299081": "This is my first Kaggle competition, so please bear with me. I made some nice notebook involving lots of pre-processing that I apply to the images, but when I copy it into a Kaggle notebook I realized that some packages are missing. To install them, I must switch on the internet. Does that mean my notebook violates the rules? Am I only allowed to use the packages that are installed by default?",
    "1299083": "No, you can use external packages. But for inference notebook you'll have to include them as Kaggle datasets for installation, since you can't submit your notebook with internet being turned on.",
    "1299618": "Thanks, I hadn't thought of that. I already found a thread explaining how to do that.",
    "1306456": "share link please",
    "1306504": "[Try this discussion](https://www.kaggle.com/samuelepino/pip-installing-packages-with-no-internet)\n\nWhat I did to successfully (well apart from my score) submit notebooks with external packages:\n\n1) Open some Kaggle notebook with internet on\n2) run `!pip download [insert package name] -d .`\n3) it will download the package and all its requirements\n4) download all this onto your computer, put it in a zip file\n5) upload it again as Kaggle data into your submission notebook\n6) load it using `!pip install [insert package name] --no-index --find-links=file:///kaggle/input/[package folder]`\n\nThere is probably a smarter way to do it but you only have to do it once per package."
  },
  "source": "meta"
}