{
  "id": 73699,
  "title": "Sketch R2CNN",
  "url": "/competitions/quickdraw-doodle-recognition/discussion/73699",
  "author_name": "Lukasz Grad",
  "post_date": "2018-12-05T00:39:54.227000",
  "votes": 11,
  "comment_count": 5,
  "views": 0,
  "content": "<p>If anyone is interested, I coded a quick example of sketch R2CNN with rasterization layer mentioned in <a href=\"https://arxiv.org/abs/1811.08170\">https://arxiv.org/abs/1811.08170</a> . It is pretty slow since it's coded in TF.</p>\n\n<p>You can view the notebook <a href=\"https://www.kaggle.com/lukeeee/sketch-r2cnn-mini\">here</a></p>\n\n<p>Will try to update it when I have more time.</p>",
  "messages": [
    {
      "id": 433293,
      "postDate": "2018-12-05T00:39:54.227Z",
      "content": "<p>If anyone is interested, I coded a quick example of sketch R2CNN with rasterization layer mentioned in <a href=\"https://arxiv.org/abs/1811.08170\">https://arxiv.org/abs/1811.08170</a> . It is pretty slow since it's coded in TF.</p>\n\n<p>You can view the notebook <a href=\"https://www.kaggle.com/lukeeee/sketch-r2cnn-mini\">here</a></p>\n\n<p>Will try to update it when I have more time.</p>",
      "rawMarkdown": "If anyone is interested, I coded a quick example of sketch R2CNN with rasterization layer mentioned in https://arxiv.org/abs/1811.08170 . It is pretty slow since it's coded in TF.\n\nYou can view the notebook [here][1]\n\nWill try to update it when I have more time.\n\n  [1]: https://www.kaggle.com/lukeeee/sketch-r2cnn-mini",
      "votes": 11
    },
    {
      "id": 435477,
      "postDate": "2018-12-08T04:43:17.597Z",
      "content": "<p>If you think this is interesting, you should check out my 8th place solution in the discussion.\nThank you very much for sharing this!</p>",
      "rawMarkdown": "If you think this is interesting, you should check out my 8th place solution in the discussion.\nThank you very much for sharing this!",
      "votes": 1,
      "replies": [
        {
          "id": 435636,
          "postDate": "2018-12-08T12:32:36.200Z",
          "content": "<p>A very nice approach, thanks for sharing too and congrats on gold medal!</p>",
          "rawMarkdown": "A very nice approach, thanks for sharing too and congrats on gold medal!"
        }
      ]
    },
    {
      "id": 433690,
      "postDate": "2018-12-05T10:58:05.640Z",
      "content": "<p>Awesome! Thanks for sharing!</p>",
      "rawMarkdown": "Awesome! Thanks for sharing!"
    },
    {
      "id": 433564,
      "postDate": "2018-12-05T07:26:50.550Z",
      "content": "<p>Thanks for this!</p>",
      "rawMarkdown": "Thanks for this!"
    },
    {
      "id": 433330,
      "postDate": "2018-12-05T01:27:19.793Z",
      "content": "<p>Thank you for the implementation!</p>",
      "rawMarkdown": "Thank you for the implementation!"
    }
  ],
  "comments": [
    {
      "id": 435477,
      "author_name": "Aleksey Nozdryn-Plotnicki",
      "author_url": "",
      "post_date": "2018-12-08T04:43:17.597000",
      "content": "<p>If you think this is interesting, you should check out my 8th place solution in the discussion.\nThank you very much for sharing this!</p>",
      "votes": 1,
      "replies": [
        {
          "id": 435636,
          "author_name": "Lukasz Grad",
          "author_url": "",
          "post_date": "2018-12-08T12:32:36.200000",
          "content": "<p>A very nice approach, thanks for sharing too and congrats on gold medal!</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 433690,
      "author_name": "[he.ai]soulmachine",
      "author_url": "",
      "post_date": "2018-12-05T10:58:05.640000",
      "content": "<p>Awesome! Thanks for sharing!</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 433564,
      "author_name": "lippyt",
      "author_url": "",
      "post_date": "2018-12-05T07:26:50.550000",
      "content": "<p>Thanks for this!</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 433330,
      "author_name": "Soonhwan Kwon",
      "author_url": "",
      "post_date": "2018-12-05T01:27:19.793000",
      "content": "<p>Thank you for the implementation!</p>",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "433293": "If anyone is interested, I coded a quick example of sketch R2CNN with rasterization layer mentioned in https://arxiv.org/abs/1811.08170 . It is pretty slow since it's coded in TF.\n\nYou can view the notebook [here][1]\n\nWill try to update it when I have more time.\n\n  [1]: https://www.kaggle.com/lukeeee/sketch-r2cnn-mini",
    "435477": "If you think this is interesting, you should check out my 8th place solution in the discussion.\nThank you very much for sharing this!",
    "433690": "Awesome! Thanks for sharing!",
    "433564": "Thanks for this!",
    "433330": "Thank you for the implementation!"
  }
}