{
  "id": 37138,
  "title": "Cloud storage hosting",
  "url": "/competitions/imagenet-object-localization-challenge/discussion/37138",
  "author_name": "RossWightman",
  "post_date": "2017-07-27T16:56:37.941000",
  "votes": 6,
  "comment_count": 0,
  "views": 0,
  "content": "<p>With this move, it would be good to see the dataset hosted in the Google cloud storage and accessible from GCE instances and the ML service. Anyone using it would still be burning through compute time but avoiding the need to host 150+GB of data under your account is one less barrier. </p>\n\n<p>Even better if we can get COCO, Imagnet 5/10/22K etc and other large datasets available.</p>",
  "messages": [
    {
      "id": 207853,
      "postDate": "2017-07-27T16:56:37.943Z",
      "content": "<p>With this move, it would be good to see the dataset hosted in the Google cloud storage and accessible from GCE instances and the ML service. Anyone using it would still be burning through compute time but avoiding the need to host 150+GB of data under your account is one less barrier. </p>\n\n<p>Even better if we can get COCO, Imagnet 5/10/22K etc and other large datasets available.</p>",
      "rawMarkdown": "With this move, it would be good to see the dataset hosted in the Google cloud storage and accessible from GCE instances and the ML service. Anyone using it would still be burning through compute time but avoiding the need to host 150+GB of data under your account is one less barrier. \n\nEven better if we can get COCO, Imagnet 5/10/22K etc and other large datasets available.",
      "votes": 5
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "207853": "With this move, it would be good to see the dataset hosted in the Google cloud storage and accessible from GCE instances and the ML service. Anyone using it would still be burning through compute time but avoiding the need to host 150+GB of data under your account is one less barrier. \n\nEven better if we can get COCO, Imagnet 5/10/22K etc and other large datasets available."
  }
}