{
  "id": 412468,
  "title": "How large should we expect about the test data?",
  "url": "/competitions/vesuvius-challenge-ink-detection/discussion/412468",
  "author_name": "cactus",
  "post_date": "2023-05-23T20:58:31.061000",
  "votes": 0,
  "comment_count": 0,
  "views": 0,
  "content": "<p>I apologize if this question was asked by anyone before but I could not find a desired post.</p>\n<p>Since the test data may be replaced during actual submission, I was wondering how many and how large should we expect about the real test data? </p>\n<p>I think this matters since I am trying to implement a dataloader that will load all the training/testing data all at once with some sort of pruning. The expected size of testing data would greatly affect the implementation we can use since we need to ensure the loaded data will fit into the memory.</p>",
  "messages": [
    {
      "id": 2271388,
      "postDate": "2023-05-23T20:58:31.063Z",
      "content": "<p>I apologize if this question was asked by anyone before but I could not find a desired post.</p>\n<p>Since the test data may be replaced during actual submission, I was wondering how many and how large should we expect about the real test data? </p>\n<p>I think this matters since I am trying to implement a dataloader that will load all the training/testing data all at once with some sort of pruning. The expected size of testing data would greatly affect the implementation we can use since we need to ensure the loaded data will fit into the memory.</p>",
      "rawMarkdown": "I apologize if this question was asked by anyone before but I could not find a desired post.\n\nSince the test data may be replaced during actual submission, I was wondering how many and how large should we expect about the real test data? \n\nI think this matters since I am trying to implement a dataloader that will load all the training/testing data all at once with some sort of pruning. The expected size of testing data would greatly affect the implementation we can use since we need to ensure the loaded data will fit into the memory."
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "2271388": "I apologize if this question was asked by anyone before but I could not find a desired post.\n\nSince the test data may be replaced during actual submission, I was wondering how many and how large should we expect about the real test data? \n\nI think this matters since I am trying to implement a dataloader that will load all the training/testing data all at once with some sort of pruning. The expected size of testing data would greatly affect the implementation we can use since we need to ensure the loaded data will fit into the memory."
  }
}