{
  "id": 10534,
  "title": "Required model documentation and code",
  "content": "<p>Hello, a team declined the prize so for me to get it I must submit my solution.</p>\n<p><a href=\"https://github.com/juliandewit/kaggle_criteo/blob/master/ModelDocumentation.pdf\" target=\"_blank\">Here is the model documentation</a></p>\n<p><a href=\"https://github.com/juliandewit/kaggle_criteo\" target=\"_blank\">Here is the code repository</a></p>\n<p>Note that the neural network was home-grown and still work-in-progress. I &nbsp;fixed bugs and added features as needed during the competition. It's also written in C# which is rather... &quot;exotic&quot; here I think :)</p>\n<p>Any feedback and tips are welcome.</p>",
  "comments": [
    {
      "id": 55683,
      "author_name": "tinrtgu",
      "author_url": "",
      "post_date": "2014-10-06T14:12:51.310000",
      "content": "<p>Congrats on getting the prize!</p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 79207,
      "author_name": "Julian de Wit",
      "author_url": "",
      "post_date": "2015-05-19T09:32:39.887000",
      "content": "<p>If it&nbsp;was that simple you would see more neural network solutions in this competition.<br>You can read the paper why some modifications were necessary.</p>\n<ol>\n<li>Caffe does not support sparsity. Some features had 100000+ categories.<br>My net had 250000 inputs. With dense&nbsp;matrices this won't work speedwise and memorywise.&nbsp;This was the reason why you did not see&nbsp;many neural nets in this competition</li>\n<li>I like to build stuff myself :)</li>\n</ol>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 79201,
      "author_name": "Chenglei Niu",
      "author_url": "",
      "post_date": "2015-05-19T09:15:04.803000",
      "content": "<p>why you don't use caffe, it allow you just need prepare the train data and network configuration file then run it you get what you want.</p>",
      "votes": 0,
      "replies": []
    }
  ],
  "comment_count": 3,
  "messages": [
    {
      "id": 55646,
      "postDate": "2014-10-05T18:33:57.773Z",
      "content": "<p>Hello, a team declined the prize so for me to get it I must submit my solution.</p>\n<p><a href=\"https://github.com/juliandewit/kaggle_criteo/blob/master/ModelDocumentation.pdf\" target=\"_blank\">Here is the model documentation</a></p>\n<p><a href=\"https://github.com/juliandewit/kaggle_criteo\" target=\"_blank\">Here is the code repository</a></p>\n<p>Note that the neural network was home-grown and still work-in-progress. I &nbsp;fixed bugs and added features as needed during the competition. It's also written in C# which is rather... &quot;exotic&quot; here I think :)</p>\n<p>Any feedback and tips are welcome.</p>",
      "votes": 11
    },
    {
      "id": 55683,
      "postDate": "2014-10-06T14:12:51.310Z",
      "content": "<p>Congrats on getting the prize!</p>",
      "votes": 1
    },
    {
      "id": 79207,
      "postDate": "2015-05-19T09:32:39.887Z",
      "content": "<p>If it&nbsp;was that simple you would see more neural network solutions in this competition.<br>You can read the paper why some modifications were necessary.</p>\n<ol>\n<li>Caffe does not support sparsity. Some features had 100000+ categories.<br>My net had 250000 inputs. With dense&nbsp;matrices this won't work speedwise and memorywise.&nbsp;This was the reason why you did not see&nbsp;many neural nets in this competition</li>\n<li>I like to build stuff myself :)</li>\n</ol>"
    },
    {
      "id": 79201,
      "postDate": "2015-05-19T09:15:04.803Z",
      "content": "<p>why you don't use caffe, it allow you just need prepare the train data and network configuration file then run it you get what you want.</p>"
    }
  ],
  "complete": true,
  "source": "live",
  "fetched_at": "2026-10-19T14:43:58Z",
  "competition_id": 3934,
  "competition_slug": "criteo-display-ad-challenge",
  "team_id": 110850,
  "solution_url": "/competitions/criteo-display-ad-challenge/writeups/julian-de-wit-required-model-documentation-and-cod"
}