{
  "id": 61218,
  "title": "And yet another request regarding the training set",
  "url": "/competitions/google-ai-open-images-object-detection-track/discussion/61218",
  "author_name": "Moshel",
  "post_date": "2018-07-16T02:20:08.411000",
  "votes": 5,
  "comment_count": 0,
  "views": 0,
  "content": "<p>As the process of converting the files to tfrecord takes forever and also uses twice (the already obnoxious) disc space requirements, it would be super duper helpful if google/kaggle can pack the files into train and validation tfrecord files (with the groundtruth) and place them somewhere nice. This will really help those of us who live in god forsaken countries with threadbare connection to the internet and have less then 2tb of disk space available....</p>",
  "messages": [
    {
      "id": 357426,
      "postDate": "2018-07-16T02:20:08.410Z",
      "content": "<p>As the process of converting the files to tfrecord takes forever and also uses twice (the already obnoxious) disc space requirements, it would be super duper helpful if google/kaggle can pack the files into train and validation tfrecord files (with the groundtruth) and place them somewhere nice. This will really help those of us who live in god forsaken countries with threadbare connection to the internet and have less then 2tb of disk space available....</p>",
      "rawMarkdown": "As the process of converting the files to tfrecord takes forever and also uses twice (the already obnoxious) disc space requirements, it would be super duper helpful if google/kaggle can pack the files into train and validation tfrecord files (with the groundtruth) and place them somewhere nice. This will really help those of us who live in god forsaken countries with threadbare connection to the internet and have less then 2tb of disk space available....",
      "votes": 4
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "357426": "As the process of converting the files to tfrecord takes forever and also uses twice (the already obnoxious) disc space requirements, it would be super duper helpful if google/kaggle can pack the files into train and validation tfrecord files (with the groundtruth) and place them somewhere nice. This will really help those of us who live in god forsaken countries with threadbare connection to the internet and have less then 2tb of disk space available...."
  }
}