{
  "id": 49064,
  "title": "Sharing new training dataset",
  "url": "/competitions/sp-society-camera-model-identification/discussion/49064",
  "author_name": "",
  "post_date": "2018-02-06T14:41:27.476968800Z",
  "votes": 16,
  "comment_count": 6,
  "views": 0,
  "content": "<p>Here's a script to download ~ 84 Gb worth of Flickr CC images:</p>\n\n<pre><code>$ du -h\n11G ./samsung_s4\n1,7G    ./nexus_5x\n9,0G    ./nexus_6\n5,7G    ./iphone_4s\n6,1G    ./moto_x\n13G ./moto_maxx\n8,4G    ./samsung_note3\n19G ./sony_nex7\n6,7G    ./iphone_6\n4,6G    ./htc_m7\n84G .\n</code></pre>\n\n<p>As I said in the other thread, you need to filter out these but as the saying goes \"my job here is done\" wrt: sharing dataset, this is a superset of the dataset I'm using, e.g. moto_x only has XT1060 model (Moto X 1st gen). </p>\n\n<p>If you use this dataset and you get more Moto-X samples, please share! :-)</p>",
  "messages": [
    {
      "id": "278598",
      "postDate": "02/06/2018 14:41:27",
      "content": "<p>Here's a script to download ~ 84 Gb worth of Flickr CC images:</p>\n\n<pre><code>$ du -h\n11G ./samsung_s4\n1,7G    ./nexus_5x\n9,0G    ./nexus_6\n5,7G    ./iphone_4s\n6,1G    ./moto_x\n13G ./moto_maxx\n8,4G    ./samsung_note3\n19G ./sony_nex7\n6,7G    ./iphone_6\n4,6G    ./htc_m7\n84G .\n</code></pre>\n\n<p>As I said in the other thread, you need to filter out these but as the saying goes \"my job here is done\" wrt: sharing dataset, this is a superset of the dataset I'm using, e.g. moto_x only has XT1060 model (Moto X 1st gen). </p>\n\n<p>If you use this dataset and you get more Moto-X samples, please share! :-)</p>",
      "rawMarkdown": "Here's a script to download ~ 84 Gb worth of Flickr CC images:\n\n    $ du -h\n    11G\t./samsung_s4\n    1,7G\t./nexus_5x\n    9,0G\t./nexus_6\n    5,7G\t./iphone_4s\n    6,1G\t./moto_x\n    13G\t./moto_maxx\n    8,4G\t./samsung_note3\n    19G\t./sony_nex7\n    6,7G\t./iphone_6\n    4,6G\t./htc_m7\n    84G\t.\n\nAs I said in the other thread, you need to filter out these but as the saying goes \"my job here is done\" wrt: sharing dataset, this is a superset of the dataset I'm using, e.g. moto_x only has XT1060 model (Moto X 1st gen). \n\nIf you use this dataset and you get more Moto-X samples, please share! :-)",
      "votes": null
    },
    {
      "id": "278664",
      "postDate": "02/06/2018 17:25:39",
      "content": "<p>Well, I think I have no time to download the data and train some new models.</p>",
      "rawMarkdown": "Well, I think I have no time to download the data and train some new models.",
      "votes": null
    },
    {
      "id": "278933",
      "postDate": "02/07/2018 04:26:28",
      "content": "<p>It's really a big data set, however I think there is no time to download it and train new models with about many GPUs. Anyway, thanks for your sharing.</p>",
      "rawMarkdown": "It's really a big data set, however I think there is no time to download it and train new models with about many GPUs. Anyway, thanks for your sharing.",
      "votes": null
    },
    {
      "id": "278959",
      "postDate": "02/07/2018 05:32:04",
      "content": "<p>Thanks for your sharing, \nBut I have no time ,  have no enough hard disk to download, and have no GPU to train.</p>",
      "rawMarkdown": "Thanks for your sharing, \nBut I have no time ,  have no enough hard disk to download, and have no GPU to train.",
      "votes": null
    },
    {
      "id": "279780",
      "postDate": "02/08/2018 17:31:59",
      "content": "<p>As much I wish to download this data, the flickr api is somehow not working for me...</p>",
      "rawMarkdown": "As much I wish to download this data, the flickr api is somehow not working for me...",
      "votes": null
    },
    {
      "id": "279927",
      "postDate": "02/09/2018 00:37:56",
      "content": "<p>Thank you once again Andres for sharing not only your approach, but even your code and this great dataset! I'm sure we've all learnt a lot from you.</p>",
      "rawMarkdown": "Thank you once again Andres for sharing not only your approach, but even your code and this great dataset! I'm sure we've all learnt a lot from you.",
      "votes": null
    },
    {
      "id": "281062",
      "postDate": "02/11/2018 21:19:49",
      "content": "<p>Thanks for sharing.</p>",
      "rawMarkdown": "Thanks for sharing.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 278664,
      "author_name": "youngkl",
      "author_url": "",
      "post_date": "02/06/2018 17:25:39",
      "content": "<p>Well, I think I have no time to download the data and train some new models.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 278933,
      "author_name": "newlifer",
      "author_url": "",
      "post_date": "02/07/2018 04:26:28",
      "content": "<p>It's really a big data set, however I think there is no time to download it and train new models with about many GPUs. Anyway, thanks for your sharing.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 278959,
      "author_name": "yyll008",
      "author_url": "",
      "post_date": "02/07/2018 05:32:04",
      "content": "<p>Thanks for your sharing, \nBut I have no time ,  have no enough hard disk to download, and have no GPU to train.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 279780,
      "author_name": "yipcma",
      "author_url": "",
      "post_date": "02/08/2018 17:31:59",
      "content": "<p>As much I wish to download this data, the flickr api is somehow not working for me...</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 279927,
      "author_name": "grassknoted",
      "author_url": "",
      "post_date": "02/09/2018 00:37:56",
      "content": "<p>Thank you once again Andres for sharing not only your approach, but even your code and this great dataset! I'm sure we've all learnt a lot from you.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 281062,
      "author_name": "kairosart",
      "author_url": "",
      "post_date": "02/11/2018 21:19:49",
      "content": "<p>Thanks for sharing.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "278598": "Here's a script to download ~ 84 Gb worth of Flickr CC images:\n\n    $ du -h\n    11G\t./samsung_s4\n    1,7G\t./nexus_5x\n    9,0G\t./nexus_6\n    5,7G\t./iphone_4s\n    6,1G\t./moto_x\n    13G\t./moto_maxx\n    8,4G\t./samsung_note3\n    19G\t./sony_nex7\n    6,7G\t./iphone_6\n    4,6G\t./htc_m7\n    84G\t.\n\nAs I said in the other thread, you need to filter out these but as the saying goes \"my job here is done\" wrt: sharing dataset, this is a superset of the dataset I'm using, e.g. moto_x only has XT1060 model (Moto X 1st gen). \n\nIf you use this dataset and you get more Moto-X samples, please share! :-)",
    "278664": "Well, I think I have no time to download the data and train some new models.",
    "278933": "It's really a big data set, however I think there is no time to download it and train new models with about many GPUs. Anyway, thanks for your sharing.",
    "278959": "Thanks for your sharing, \nBut I have no time ,  have no enough hard disk to download, and have no GPU to train.",
    "279780": "As much I wish to download this data, the flickr api is somehow not working for me...",
    "279927": "Thank you once again Andres for sharing not only your approach, but even your code and this great dataset! I'm sure we've all learnt a lot from you.",
    "281062": "Thanks for sharing."
  },
  "source": "meta"
}