{
  "id": 71309,
  "title": "baseball or baseball bat?",
  "url": "/competitions/quickdraw-doodle-recognition/discussion/71309",
  "author_name": "hengck23",
  "post_date": "2018-11-12T14:00:44.401000",
  "votes": 6,
  "comment_count": 3,
  "views": 0,
  "content": "<p>the pause time gives the clue, see below</p>\n\n<p><img src=\"https://storage.googleapis.com/kaggle-forum-message-attachments/419748/10660/pause_time.png\" alt=\"enter image description here\"></p>",
  "messages": [
    {
      "id": 419748,
      "postDate": "2018-11-12T14:00:44.400Z",
      "content": "<p>the pause time gives the clue, see below</p>\n\n<p><img src=\"https://storage.googleapis.com/kaggle-forum-message-attachments/419748/10660/pause_time.png\" alt=\"enter image description here\"></p>",
      "rawMarkdown": "the pause time gives the clue, see below\n\n\n  ![enter image description here][1]\n\n\n  [1]: https://storage.googleapis.com/kaggle-forum-message-attachments/419748/10660/pause_time.png",
      "votes": 6
    },
    {
      "id": 421027,
      "postDate": "2018-11-14T13:48:07.293Z",
      "content": "<p>after hand-cleaning up the non-recognized labels, i have become expert in quickdraw :</p>\n\n<p>some strokes are more important than the rest.\ncreate an attention lstm on the strokes  and visualize the attention weights will reveal it.</p>\n\n<p><img src=\"https://storage.googleapis.com/kaggle-forum-message-attachments/421027/10676/6_correct.png\" alt=\"enter image description here\"></p>",
      "rawMarkdown": "after hand-cleaning up the non-recognized labels, i have become expert in quickdraw :\n\nsome strokes are more important than the rest.\ncreate an attention lstm on the strokes  and visualize the attention weights will reveal it.\n\n   ![enter image description here][1]\n\n\n  [1]: https://storage.googleapis.com/kaggle-forum-message-attachments/421027/10676/6_correct.png",
      "votes": 4
    },
    {
      "id": 422503,
      "postDate": "2018-11-16T10:18:36.423Z",
      "content": "<p>This is so cool and amazing, waiting to learn a lot more from you note.. A quick question may be naive is, should there always be a encoder-decoder architecture for attention mechanism implementation..</p>",
      "rawMarkdown": "This is so cool and amazing, waiting to learn a lot more from you note.. A quick question may be naive is, should there always be a encoder-decoder architecture for attention mechanism implementation.."
    },
    {
      "id": 421036,
      "postDate": "2018-11-14T13:57:45.013Z",
      "content": "<p>more here</p>\n\n<p><img src=\"https://storage.googleapis.com/kaggle-forum-message-attachments/421036/10677/6c1.png\" alt=\"enter image description here\"></p>",
      "rawMarkdown": "more here\n\n\n  ![enter image description here][1]\n\n\n  [1]: https://storage.googleapis.com/kaggle-forum-message-attachments/421036/10677/6c1.png"
    }
  ],
  "comments": [
    {
      "id": 421027,
      "author_name": "hengck23",
      "author_url": "",
      "post_date": "2018-11-14T13:48:07.293000",
      "content": "<p>after hand-cleaning up the non-recognized labels, i have become expert in quickdraw :</p>\n\n<p>some strokes are more important than the rest.\ncreate an attention lstm on the strokes  and visualize the attention weights will reveal it.</p>\n\n<p><img src=\"https://storage.googleapis.com/kaggle-forum-message-attachments/421027/10676/6_correct.png\" alt=\"enter image description here\"></p>",
      "votes": 4,
      "replies": []
    },
    {
      "id": 422503,
      "author_name": "Vishy",
      "author_url": "",
      "post_date": "2018-11-16T10:18:36.423000",
      "content": "<p>This is so cool and amazing, waiting to learn a lot more from you note.. A quick question may be naive is, should there always be a encoder-decoder architecture for attention mechanism implementation..</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 421036,
      "author_name": "hengck23",
      "author_url": "",
      "post_date": "2018-11-14T13:57:45.013000",
      "content": "<p>more here</p>\n\n<p><img src=\"https://storage.googleapis.com/kaggle-forum-message-attachments/421036/10677/6c1.png\" alt=\"enter image description here\"></p>",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "419748": "the pause time gives the clue, see below\n\n\n  ![enter image description here][1]\n\n\n  [1]: https://storage.googleapis.com/kaggle-forum-message-attachments/419748/10660/pause_time.png",
    "421027": "after hand-cleaning up the non-recognized labels, i have become expert in quickdraw :\n\nsome strokes are more important than the rest.\ncreate an attention lstm on the strokes  and visualize the attention weights will reveal it.\n\n   ![enter image description here][1]\n\n\n  [1]: https://storage.googleapis.com/kaggle-forum-message-attachments/421027/10676/6_correct.png",
    "422503": "This is so cool and amazing, waiting to learn a lot more from you note.. A quick question may be naive is, should there always be a encoder-decoder architecture for attention mechanism implementation..",
    "421036": "more here\n\n\n  ![enter image description here][1]\n\n\n  [1]: https://storage.googleapis.com/kaggle-forum-message-attachments/421036/10677/6c1.png"
  }
}