{
  "id": 432879,
  "title": "Another Pre-processed false-color image dataset",
  "url": "/competitions/google-research-identify-contrails-reduce-global-warming/discussion/432879",
  "author_name": "Bilzard",
  "post_date": "2023-08-19T10:00:31.804000",
  "votes": 2,
  "comment_count": 0,
  "views": 0,
  "content": "<p>For the people who wants to test their model in late submission, I published the pre-processed false-color image dataset I used during the competition.</p>\n<p>There are three type of datasets:</p>\n<ol>\n<li><code>images</code>: false-color images+labels for all frames in the train/validation set (Data format: np.uint8)</li>\n<li><code>images_fp16</code>: false color images+labels for the current frames only in the train/validation set (Data format: np.float16)</li>\n<li><code>images_f5</code>: false-color images+labels for the first five frames (i.e. frame 0-4) in the train/validation set (Data format: np.uint8)</li>\n</ol>\n<p>[1] contains the first two, and [2] contains the latter.</p>\n<h2>Reference</h2>\n<ul>\n<li>[1] <a href=\"https://www.kaggle.com/datasets/tatamikenn/contrail-preprocessed-images\" target=\"_blank\">Contrail: Preprocessed Images</a></li>\n<li>[2] <a href=\"https://www.kaggle.com/datasets/tatamikenn/contrail-false-color-5\" target=\"_blank\">Contrail: false-color image 5 frames</a></li>\n</ul>",
  "messages": [
    {
      "id": 2397903,
      "postDate": "2023-08-19T10:00:31.803Z",
      "content": "<p>For the people who wants to test their model in late submission, I published the pre-processed false-color image dataset I used during the competition.</p>\n<p>There are three type of datasets:</p>\n<ol>\n<li><code>images</code>: false-color images+labels for all frames in the train/validation set (Data format: np.uint8)</li>\n<li><code>images_fp16</code>: false color images+labels for the current frames only in the train/validation set (Data format: np.float16)</li>\n<li><code>images_f5</code>: false-color images+labels for the first five frames (i.e. frame 0-4) in the train/validation set (Data format: np.uint8)</li>\n</ol>\n<p>[1] contains the first two, and [2] contains the latter.</p>\n<h2>Reference</h2>\n<ul>\n<li>[1] <a href=\"https://www.kaggle.com/datasets/tatamikenn/contrail-preprocessed-images\" target=\"_blank\">Contrail: Preprocessed Images</a></li>\n<li>[2] <a href=\"https://www.kaggle.com/datasets/tatamikenn/contrail-false-color-5\" target=\"_blank\">Contrail: false-color image 5 frames</a></li>\n</ul>",
      "rawMarkdown": "For the people who wants to test their model in late submission, I published the pre-processed false-color image dataset I used during the competition.\n\nThere are three type of datasets:\n\n1. `images`: false-color images+labels for all frames in the train/validation set (Data format: np.uint8)\n2. `images_fp16`: false color images+labels for the current frames only in the train/validation set (Data format: np.float16)\n3. `images_f5`: false-color images+labels for the first five frames (i.e. frame 0-4) in the train/validation set (Data format: np.uint8)\n\n[1] contains the first two, and [2] contains the latter.\n\n## Reference\n\n- [1] [Contrail: Preprocessed Images](https://www.kaggle.com/datasets/tatamikenn/contrail-preprocessed-images)\n- [2] [Contrail: false-color image 5 frames](https://www.kaggle.com/datasets/tatamikenn/contrail-false-color-5)",
      "votes": 2
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "2397903": "For the people who wants to test their model in late submission, I published the pre-processed false-color image dataset I used during the competition.\n\nThere are three type of datasets:\n\n1. `images`: false-color images+labels for all frames in the train/validation set (Data format: np.uint8)\n2. `images_fp16`: false color images+labels for the current frames only in the train/validation set (Data format: np.float16)\n3. `images_f5`: false-color images+labels for the first five frames (i.e. frame 0-4) in the train/validation set (Data format: np.uint8)\n\n[1] contains the first two, and [2] contains the latter.\n\n## Reference\n\n- [1] [Contrail: Preprocessed Images](https://www.kaggle.com/datasets/tatamikenn/contrail-preprocessed-images)\n- [2] [Contrail: false-color image 5 frames](https://www.kaggle.com/datasets/tatamikenn/contrail-false-color-5)"
  }
}