{
  "id": 353399,
  "title": "Question About Validation Set - Representative, Number of Folds/Split",
  "url": "/competitions/tpu-getting-started/discussion/353399",
  "author_name": "KaiquanMah",
  "post_date": "2022-09-18T07:51:26.612000",
  "votes": 0,
  "comment_count": 0,
  "views": null,
  "content": "<p>From the '<a href=\"https://www.kaggle.com/competitions/tpu-getting-started/data\" target=\"_blank\">Data</a>' tab, I see that the validation set has the following description:</p>\n<blockquote>\n  <p>val/*.tfrec - <strong>pre-split training samples</strong> w/ labels intended to help with checking your model's performance on TPU. <strong>The split was stratified across labels.</strong></p>\n</blockquote>\n<p>Essentially it means that</p>\n<ol>\n<li>the percentage of each label relative to all labels in the validation set is representative of the training and test set/read world?</li>\n<li>the validation set has not been split across the number of folds we want -&gt; we need to determine the 'k' number of folds to split during cross-validation ourselves??</li>\n</ol>",
  "messages": [
    {
      "id": 1944311,
      "postDate": "2022-09-18T07:51:26.613Z",
      "content": "<p>From the '<a href=\"https://www.kaggle.com/competitions/tpu-getting-started/data\" target=\"_blank\">Data</a>' tab, I see that the validation set has the following description:</p>\n<blockquote>\n  <p>val/*.tfrec - <strong>pre-split training samples</strong> w/ labels intended to help with checking your model's performance on TPU. <strong>The split was stratified across labels.</strong></p>\n</blockquote>\n<p>Essentially it means that</p>\n<ol>\n<li>the percentage of each label relative to all labels in the validation set is representative of the training and test set/read world?</li>\n<li>the validation set has not been split across the number of folds we want -&gt; we need to determine the 'k' number of folds to split during cross-validation ourselves??</li>\n</ol>",
      "rawMarkdown": "From the '[Data](https://www.kaggle.com/competitions/tpu-getting-started/data)' tab, I see that the validation set has the following description:\n> val/*.tfrec - **pre-split training samples** w/ labels intended to help with checking your model's performance on TPU. **The split was stratified across labels.**\n\nEssentially it means that\n1. the percentage of each label relative to all labels in the validation set is representative of the training and test set/read world?\n2. the validation set has not been split across the number of folds we want -> we need to determine the 'k' number of folds to split during cross-validation ourselves??"
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1944311": "From the '[Data](https://www.kaggle.com/competitions/tpu-getting-started/data)' tab, I see that the validation set has the following description:\n> val/*.tfrec - **pre-split training samples** w/ labels intended to help with checking your model's performance on TPU. **The split was stratified across labels.**\n\nEssentially it means that\n1. the percentage of each label relative to all labels in the validation set is representative of the training and test set/read world?\n2. the validation set has not been split across the number of folds we want -> we need to determine the 'k' number of folds to split during cross-validation ourselves??"
  }
}