{
  "id": 37105,
  "title": "Can you please reduce \"Data\" its about 154 GB",
  "url": "/competitions/imagenet-object-localization-challenge/discussion/37105",
  "author_name": "",
  "post_date": "2017-07-27T05:11:11.764987600Z",
  "votes": 11,
  "comment_count": 8,
  "views": 0,
  "content": "<p>please reduce the size of data to say 200 MB. I just want to see images to test my code &amp; then apply for this competition, currently I know nothing about how to make it train.. </p>",
  "messages": [
    {
      "id": "207659",
      "postDate": "07/27/2017 05:11:11",
      "content": "<p>please reduce the size of data to say 200 MB. I just want to see images to test my code &amp; then apply for this competition, currently I know nothing about how to make it train.. </p>",
      "rawMarkdown": "please reduce the size of data to say 200 MB. I just want to see images to test my code &amp; then apply for this competition, currently I know nothing about how to make it train..",
      "votes": null
    },
    {
      "id": "207882",
      "postDate": "07/27/2017 19:44:05",
      "content": "<p>There are reduced image size variants of Imagenet (see <a href=\"https://tiny-imagenet.herokuapp.com/\">https://tiny-imagenet.herokuapp.com/</a> and <a href=\"https://patrykchrabaszcz.github.io/Imagenet32/\">https://patrykchrabaszcz.github.io/Imagenet32/</a>) but the models trained with them will obviously not be comparable to those trained on the larger images in terms of performance or feature scale. </p>\n\n<p>It looks like the scaled down version reference in the second link has a hosted version on the main Imagenet download page (signup required): <a href=\"http://image-net.org/download-images\">http://image-net.org/download-images</a></p>",
      "rawMarkdown": "There are reduced image size variants of Imagenet (see https://tiny-imagenet.herokuapp.com/ and https://patrykchrabaszcz.github.io/Imagenet32/) but the models trained with them will obviously not be comparable to those trained on the larger images in terms of performance or feature scale. \n\nIt looks like the scaled down version reference in the second link has a hosted version on the main Imagenet download page (signup required): http://image-net.org/download-images",
      "votes": null
    },
    {
      "id": "208932",
      "postDate": "07/31/2017 15:32:43",
      "content": "<p>I think admin should consider splitting the file into smaller ones for us to download.</p>",
      "rawMarkdown": "I think admin should consider splitting the file into smaller ones for us to download.",
      "votes": null
    },
    {
      "id": "208961",
      "postDate": "07/31/2017 17:52:19",
      "content": "<p>Please split the file or provide torrent for this data set. </p>",
      "rawMarkdown": "Please split the file or provide torrent for this data set.",
      "votes": null
    },
    {
      "id": "209158",
      "postDate": "08/01/2017 12:20:39",
      "content": "<p>Thaks so much lets make some hands dirty on it, i know nothing about it :)</p>",
      "rawMarkdown": "Thaks so much lets make some hands dirty on it, i know nothing about it :)",
      "votes": null
    },
    {
      "id": "214578",
      "postDate": "08/17/2017 12:58:32",
      "content": "<p>It takes several hours to download,then where to store so big data set is another problem.</p>",
      "rawMarkdown": "It takes several hours to download,then where to store so big data set is another problem.",
      "votes": null
    },
    {
      "id": "347967",
      "postDate": "06/25/2018 19:41:47",
      "content": "<p>A very simple and quick solution would be separating the training and test data in two files. (They will keep big, but significantly smaller)</p>",
      "rawMarkdown": "A very simple and quick solution would be separating the training and test data in two files. (They will keep big, but significantly smaller)",
      "votes": null
    },
    {
      "id": "348046",
      "postDate": "06/26/2018 01:04:42",
      "content": "<p>yeah  the data is to big and my compute cannot endure that pressure</p>",
      "rawMarkdown": "yeah  the data is to big and my compute cannot endure that pressure",
      "votes": null
    },
    {
      "id": "348284",
      "postDate": "06/26/2018 13:00:49",
      "content": "<p>I have the same problem too. It is difficult to download the data. It needs many hours to set it to the computer.</p>",
      "rawMarkdown": "I have the same problem too. It is difficult to download the data. It needs many hours to set it to the computer.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 207882,
      "author_name": "rwightman",
      "author_url": "",
      "post_date": "07/27/2017 19:44:05",
      "content": "<p>There are reduced image size variants of Imagenet (see <a href=\"https://tiny-imagenet.herokuapp.com/\">https://tiny-imagenet.herokuapp.com/</a> and <a href=\"https://patrykchrabaszcz.github.io/Imagenet32/\">https://patrykchrabaszcz.github.io/Imagenet32/</a>) but the models trained with them will obviously not be comparable to those trained on the larger images in terms of performance or feature scale. </p>\n\n<p>It looks like the scaled down version reference in the second link has a hosted version on the main Imagenet download page (signup required): <a href=\"http://image-net.org/download-images\">http://image-net.org/download-images</a></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 208932,
      "author_name": "kitman0804",
      "author_url": "",
      "post_date": "07/31/2017 15:32:43",
      "content": "<p>I think admin should consider splitting the file into smaller ones for us to download.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 208961,
      "author_name": "belgasoft",
      "author_url": "",
      "post_date": "07/31/2017 17:52:19",
      "content": "<p>Please split the file or provide torrent for this data set. </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 209158,
      "author_name": "prasadkolhatkar",
      "author_url": "",
      "post_date": "08/01/2017 12:20:39",
      "content": "<p>Thaks so much lets make some hands dirty on it, i know nothing about it :)</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 214578,
      "author_name": "likehuaer",
      "author_url": "",
      "post_date": "08/17/2017 12:58:32",
      "content": "<p>It takes several hours to download,then where to store so big data set is another problem.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 347967,
      "author_name": "danmoller",
      "author_url": "",
      "post_date": "06/25/2018 19:41:47",
      "content": "<p>A very simple and quick solution would be separating the training and test data in two files. (They will keep big, but significantly smaller)</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 348046,
      "author_name": "cableyang123",
      "author_url": "",
      "post_date": "06/26/2018 01:04:42",
      "content": "<p>yeah  the data is to big and my compute cannot endure that pressure</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 348284,
      "author_name": "anthi1984",
      "author_url": "",
      "post_date": "06/26/2018 13:00:49",
      "content": "<p>I have the same problem too. It is difficult to download the data. It needs many hours to set it to the computer.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "207659": "please reduce the size of data to say 200 MB. I just want to see images to test my code &amp; then apply for this competition, currently I know nothing about how to make it train..",
    "207882": "There are reduced image size variants of Imagenet (see https://tiny-imagenet.herokuapp.com/ and https://patrykchrabaszcz.github.io/Imagenet32/) but the models trained with them will obviously not be comparable to those trained on the larger images in terms of performance or feature scale. \n\nIt looks like the scaled down version reference in the second link has a hosted version on the main Imagenet download page (signup required): http://image-net.org/download-images",
    "208932": "I think admin should consider splitting the file into smaller ones for us to download.",
    "208961": "Please split the file or provide torrent for this data set.",
    "209158": "Thaks so much lets make some hands dirty on it, i know nothing about it :)",
    "214578": "It takes several hours to download,then where to store so big data set is another problem.",
    "347967": "A very simple and quick solution would be separating the training and test data in two files. (They will keep big, but significantly smaller)",
    "348046": "yeah  the data is to big and my compute cannot endure that pressure",
    "348284": "I have the same problem too. It is difficult to download the data. It needs many hours to set it to the computer."
  },
  "source": "meta"
}