{
  "id": 13293,
  "title": "Auroraxie's report and code",
  "url": "/competitions/datasciencebowl/writeups/auroraxie-auroraxie-s-report-and-code",
  "author_name": "",
  "post_date": "2015-04-09T08:14:10.217Z",
  "votes": 6,
  "comment_count": 3,
  "views": 1819,
  "content": "<p>Thanks Kaggle for holding this competition! Congratulations to all the winners, and thanks all the players! It's a great fun to participate this competition with all of you.</p>\n\n<p>Here is the link for my report and code:</p>\n<p>https://github.com/auroraxie/Kaggle-NDSB</p>",
  "messages": [
    {
      "id": "70100",
      "postDate": "04/09/2015 08:14:10",
      "content": "<p>Thanks Kaggle for holding this competition! Congratulations to all the winners, and thanks all the players! It's a great fun to participate this competition with all of you.</p>\n\n<p>Here is the link for my report and code:</p>\n<p>https://github.com/auroraxie/Kaggle-NDSB</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "70143",
      "postDate": "04/09/2015 19:28:18",
      "content": "<p>AuroraXie,</p>\n<p>Thanks for posting your configurations. First I though I am seeing double with so many convolution layers but maybe that was my mistake. I have assumed with 5 conv layers I could reach a good logloss but maybe it needed much more like yours: 16-17 convolution layers. What was typical iteration count for you to reach under 0.6-0.7 on GTX 980?</p>\n<p>have you ever get a good results with 6 or less conv layers?</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "71246",
      "postDate": "04/10/2015 20:02:28",
      "content": "<p>Thanks for your interest!</p>\n<p>A lot of papers show the advantage of deep convolution networks over shallow nets. &nbsp;You can see the paper by VGG (http://arxiv.org/pdf/1409.1556.pdf) as a reference.&nbsp;Deeper nets have better performance.</p>\n<p>The number of iterations I used ranged from 250-500. You can check it out in my conf files.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "71666",
      "postDate": "04/15/2015 07:56:45",
      "content": "<p>Thanks for posting, may I know how to generate the image mean.bin ?</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 70143,
      "author_name": "esorar",
      "author_url": "",
      "post_date": "04/09/2015 19:28:18",
      "content": "<p>AuroraXie,</p>\n<p>Thanks for posting your configurations. First I though I am seeing double with so many convolution layers but maybe that was my mistake. I have assumed with 5 conv layers I could reach a good logloss but maybe it needed much more like yours: 16-17 convolution layers. What was typical iteration count for you to reach under 0.6-0.7 on GTX 980?</p>\n<p>have you ever get a good results with 6 or less conv layers?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 71246,
      "author_name": "auroraxie",
      "author_url": "",
      "post_date": "04/10/2015 20:02:28",
      "content": "<p>Thanks for your interest!</p>\n<p>A lot of papers show the advantage of deep convolution networks over shallow nets. &nbsp;You can see the paper by VGG (http://arxiv.org/pdf/1409.1556.pdf) as a reference.&nbsp;Deeper nets have better performance.</p>\n<p>The number of iterations I used ranged from 250-500. You can check it out in my conf files.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 71666,
      "author_name": "stevendu",
      "author_url": "",
      "post_date": "04/15/2015 07:56:45",
      "content": "<p>Thanks for posting, may I know how to generate the image mean.bin ?</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "70100": "",
    "70143": "",
    "71246": "",
    "71666": ""
  },
  "source": "existing_discussion",
  "complete": true,
  "competition_id": 3978,
  "competition_slug": "datasciencebowl",
  "team_id": 134789,
  "solution_url": "/competitions/datasciencebowl/writeups/auroraxie-auroraxie-s-report-and-code"
}