{
  "id": 341033,
  "title": "machine resources",
  "url": "/competitions/amex-default-prediction/discussion/341033",
  "author_name": "",
  "post_date": "2022-08-01T01:27:06.660002900Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Hi,</p>\n<p>Is it required for the final submission to be computed out of the completion data entirely on Kaggle with the provided 13GB+GPU or 16G CPU only machine?<br>\nI am new to Kaggle competitions and as the amount of data provided by AMEX is huge it is tempting to use a machine with more resources to at least prepare the data, upload it to Kaggle and learn/predict from there.</p>\n<p>thanks<br>\nNorbert </p>",
  "messages": [
    {
      "id": "1879297",
      "postDate": "08/01/2022 01:27:06",
      "content": "<p>Hi,</p>\n<p>Is it required for the final submission to be computed out of the completion data entirely on Kaggle with the provided 13GB+GPU or 16G CPU only machine?<br>\nI am new to Kaggle competitions and as the amount of data provided by AMEX is huge it is tempting to use a machine with more resources to at least prepare the data, upload it to Kaggle and learn/predict from there.</p>\n<p>thanks<br>\nNorbert </p>",
      "rawMarkdown": "Hi,\n\nIs it required for the final submission to be computed out of the completion data entirely on Kaggle with the provided 13GB+GPU or 16G CPU only machine?\nI am new to Kaggle competitions and as the amount of data provided by AMEX is huge it is tempting to use a machine with more resources to at least prepare the data, upload it to Kaggle and learn/predict from there.\n\nthanks\nNorbert",
      "votes": null
    },
    {
      "id": "1879309",
      "postDate": "08/01/2022 01:56:39",
      "content": "<p>In Kaggle \"code competitions\", predictions for submission must be done in the Kaggle online environment, but this AMEX contest is not one of those. We (I speak unofficially as a contestant, not an organizer) are free to preprocess the data, train our models, and generate our predictions on our own machines if we choose.</p>",
      "rawMarkdown": "In Kaggle \"code competitions\", predictions for submission must be done in the Kaggle online environment, but this AMEX contest is not one of those. We (I speak unofficially as a contestant, not an organizer) are free to preprocess the data, train our models, and generate our predictions on our own machines if we choose.",
      "votes": null
    },
    {
      "id": "1879501",
      "postDate": "08/01/2022 05:13:07",
      "content": "<p>thanks I did not know this so far</p>",
      "rawMarkdown": "thanks I did not know this so far",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1879309,
      "author_name": "dslate",
      "author_url": "",
      "post_date": "08/01/2022 01:56:39",
      "content": "<p>In Kaggle \"code competitions\", predictions for submission must be done in the Kaggle online environment, but this AMEX contest is not one of those. We (I speak unofficially as a contestant, not an organizer) are free to preprocess the data, train our models, and generate our predictions on our own machines if we choose.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1879501,
      "author_name": "nhenseler",
      "author_url": "",
      "post_date": "08/01/2022 05:13:07",
      "content": "<p>thanks I did not know this so far</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1879297": "Hi,\n\nIs it required for the final submission to be computed out of the completion data entirely on Kaggle with the provided 13GB+GPU or 16G CPU only machine?\nI am new to Kaggle competitions and as the amount of data provided by AMEX is huge it is tempting to use a machine with more resources to at least prepare the data, upload it to Kaggle and learn/predict from there.\n\nthanks\nNorbert",
    "1879309": "In Kaggle \"code competitions\", predictions for submission must be done in the Kaggle online environment, but this AMEX contest is not one of those. We (I speak unofficially as a contestant, not an organizer) are free to preprocess the data, train our models, and generate our predictions on our own machines if we choose.",
    "1879501": "thanks I did not know this so far"
  },
  "source": "meta"
}