{
  "id": 442859,
  "title": "Not allowed to use internet then how to install bnunicodenormalizer and jiwer?",
  "url": "/competitions/bengaliai-speech/discussion/442859",
  "author_name": "",
  "post_date": "2023-09-24T13:03:45.304469700Z",
  "votes": 1,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Title.</p>\n<p>These are needed for this competition right?</p>",
  "messages": [
    {
      "id": "2453967",
      "postDate": "09/24/2023 13:03:45",
      "content": "<p>Title.</p>\n<p>These are needed for this competition right?</p>",
      "rawMarkdown": "Title.\n\nThese are needed for this competition right?",
      "votes": null
    },
    {
      "id": "2453979",
      "postDate": "09/24/2023 13:11:13",
      "content": "<p>Not necessarily needed but they are useful. You can create a dataset like this with the wheel file, add it to your notebook, and just run <code>pip install your_wheel_file</code></p>\n<p><a href=\"https://www.kaggle.com/datasets/snnclsr/jiwer-3-0-3\" target=\"_blank\">https://www.kaggle.com/datasets/snnclsr/jiwer-3-0-3</a></p>",
      "rawMarkdown": "Not necessarily needed but they are useful. You can create a dataset like this with the wheel file, add it to your notebook, and just run `pip install your_wheel_file`\n\nhttps://www.kaggle.com/datasets/snnclsr/jiwer-3-0-3",
      "votes": null
    },
    {
      "id": "2454045",
      "postDate": "09/24/2023 14:22:00",
      "content": "<p>yeah. I just found a post about it. It seems I can add whatever through add data feature</p>",
      "rawMarkdown": "yeah. I just found a post about it. It seems I can add whatever through add data feature",
      "votes": null
    },
    {
      "id": "2454665",
      "postDate": "09/25/2023 02:38:16",
      "content": "<p>You may download and install these packages using another kernel and save the result as a dataset. Then you may use the other kernel itself/ saved dataset as input to your inference kernel and go ahead <a href=\"https://www.kaggle.com/sayemprodhanananta\" target=\"_blank\">@sayemprodhanananta</a> </p>",
      "rawMarkdown": "You may download and install these packages using another kernel and save the result as a dataset. Then you may use the other kernel itself/ saved dataset as input to your inference kernel and go ahead @sayemprodhanananta",
      "votes": null
    },
    {
      "id": "2460008",
      "postDate": "09/28/2023 14:22:08",
      "content": "<p>This approach works faster and is simpler than collecting the wheel files. Are there any downsides to this, considering the notebook environment is pinned?</p>\n<p><code>!pip install somepackage==1.0.0 --no-deps --target ./target_dir</code></p>\n<pre><code> sys\nsys.path.insert(,)\n</code></pre>",
      "rawMarkdown": "This approach works faster and is simpler than collecting the wheel files. Are there any downsides to this, considering the notebook environment is pinned?\n\n`!pip install somepackage==1.0.0 --no-deps --target ./target_dir`\n\n```python\nimport sys\nsys.path.insert(1,'/kaggle/input/target_dir/')\n```",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2453979,
      "author_name": "snnclsr",
      "author_url": "",
      "post_date": "09/24/2023 13:11:13",
      "content": "<p>Not necessarily needed but they are useful. You can create a dataset like this with the wheel file, add it to your notebook, and just run <code>pip install your_wheel_file</code></p>\n<p><a href=\"https://www.kaggle.com/datasets/snnclsr/jiwer-3-0-3\" target=\"_blank\">https://www.kaggle.com/datasets/snnclsr/jiwer-3-0-3</a></p>",
      "votes": null,
      "replies": [
        {
          "id": 2454045,
          "author_name": "sayemprodhanananta",
          "author_url": "",
          "post_date": "09/24/2023 14:22:00",
          "content": "<p>yeah. I just found a post about it. It seems I can add whatever through add data feature</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 2460008,
          "author_name": "tsobolev",
          "author_url": "",
          "post_date": "09/28/2023 14:22:08",
          "content": "<p>This approach works faster and is simpler than collecting the wheel files. Are there any downsides to this, considering the notebook environment is pinned?</p>\n<p><code>!pip install somepackage==1.0.0 --no-deps --target ./target_dir</code></p>\n<pre><code> sys\nsys.path.insert(,)\n</code></pre>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 2454665,
      "author_name": "ravi20076",
      "author_url": "",
      "post_date": "09/25/2023 02:38:16",
      "content": "<p>You may download and install these packages using another kernel and save the result as a dataset. Then you may use the other kernel itself/ saved dataset as input to your inference kernel and go ahead <a href=\"https://www.kaggle.com/sayemprodhanananta\" target=\"_blank\">@sayemprodhanananta</a> </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2453967": "Title.\n\nThese are needed for this competition right?",
    "2453979": "Not necessarily needed but they are useful. You can create a dataset like this with the wheel file, add it to your notebook, and just run `pip install your_wheel_file`\n\nhttps://www.kaggle.com/datasets/snnclsr/jiwer-3-0-3",
    "2454045": "yeah. I just found a post about it. It seems I can add whatever through add data feature",
    "2454665": "You may download and install these packages using another kernel and save the result as a dataset. Then you may use the other kernel itself/ saved dataset as input to your inference kernel and go ahead @sayemprodhanananta",
    "2460008": "This approach works faster and is simpler than collecting the wheel files. Are there any downsides to this, considering the notebook environment is pinned?\n\n`!pip install somepackage==1.0.0 --no-deps --target ./target_dir`\n\n```python\nimport sys\nsys.path.insert(1,'/kaggle/input/target_dir/')\n```"
  },
  "source": "meta"
}