{
  "id": 72553,
  "title": "Join a team (LB 935)",
  "url": "/competitions/quickdraw-doodle-recognition/discussion/72553",
  "author_name": "Eric Bouteillon",
  "post_date": "2018-11-24T13:08:02.073000",
  "votes": 3,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Hi Kagglers,</p>\n\n<p>I would like to join a team for this competition. As team deadline is close, it's really urgent for me to ask this to you. 😃</p>\n\n<p>What I have so far:</p>\n\n<ul>\n<li><p>A single CNN model trained on 50k/class, no TTA, no ensemble and giving LB 928. So could be potentially improved.</p></li>\n<li><p>A recent GPU</p></li>\n<li><p>I am bad at drawing doodles</p></li>\n</ul>\n\n<p>I would also be grateful to learn from more Kaggle experienced teammates. 🙏</p>\n\n<p>Looking forward teaming up with you.</p>\n\n<hr>\n\n<p>Update with full dataset : LB 935</p>",
  "messages": [
    {
      "id": 427057,
      "postDate": "2018-11-24T13:08:02.073Z",
      "content": "<p>Hi Kagglers,</p>\n\n<p>I would like to join a team for this competition. As team deadline is close, it's really urgent for me to ask this to you. 😃</p>\n\n<p>What I have so far:</p>\n\n<ul>\n<li><p>A single CNN model trained on 50k/class, no TTA, no ensemble and giving LB 928. So could be potentially improved.</p></li>\n<li><p>A recent GPU</p></li>\n<li><p>I am bad at drawing doodles</p></li>\n</ul>\n\n<p>I would also be grateful to learn from more Kaggle experienced teammates. 🙏</p>\n\n<p>Looking forward teaming up with you.</p>\n\n<hr>\n\n<p>Update with full dataset : LB 935</p>",
      "rawMarkdown": "Hi Kagglers,\n\nI would like to join a team for this competition. As team deadline is close, it's really urgent for me to ask this to you. 😃\n\nWhat I have so far:\n\n - A single CNN model trained on 50k/class, no TTA, no ensemble and giving LB 928. So could be potentially improved.\n\n - A recent GPU\n\n - I am bad at drawing doodles\n\nI would also be grateful to learn from more Kaggle experienced teammates. 🙏\n\nLooking forward teaming up with you.\n\n***\nUpdate with full dataset : LB 935\n",
      "votes": 3
    },
    {
      "id": 427870,
      "postDate": "2018-11-26T10:07:03.723Z",
      "content": "<p>Hi  Bouteillon,\nI want to join your team and be grateful to learn from your experience. :)\nWhat I have so far:\n 1. Two CNN models of inceptionv3 and resent50 trained on 30k/class, no TTA, no ensemble and giving LB 935 and 932 respectively.\n 2. Two 1080Ti GPU\n 3. Trying to build the right RNN network. At this stage, there is only one LSTM solution that is LB 800 on 30k/class, no TTA, no ensemble </p>\n\n<p>I have made a merge request. Looking forward teaming up with you.:)</p>",
      "rawMarkdown": "Hi  Bouteillon,\nI want to join your team and be grateful to learn from your experience. :)\nWhat I have so far:\n 1. Two CNN models of inceptionv3 and resent50 trained on 30k/class, no TTA, no ensemble and giving LB 935 and 932 respectively.\n 2. Two 1080Ti GPU\n 3. Trying to build the right RNN network. At this stage, there is only one LSTM solution that is LB 800 on 30k/class, no TTA, no ensemble \n\nI have made a merge request. Looking forward teaming up with you.:)"
    },
    {
      "id": 427180,
      "postDate": "2018-11-24T20:13:29.657Z",
      "content": "<p>I want a team tooo :)</p>",
      "rawMarkdown": "I want a team tooo :)"
    }
  ],
  "comments": [
    {
      "id": 427870,
      "author_name": "kalili",
      "author_url": "",
      "post_date": "2018-11-26T10:07:03.723000",
      "content": "<p>Hi  Bouteillon,\nI want to join your team and be grateful to learn from your experience. :)\nWhat I have so far:\n 1. Two CNN models of inceptionv3 and resent50 trained on 30k/class, no TTA, no ensemble and giving LB 935 and 932 respectively.\n 2. Two 1080Ti GPU\n 3. Trying to build the right RNN network. At this stage, there is only one LSTM solution that is LB 800 on 30k/class, no TTA, no ensemble </p>\n\n<p>I have made a merge request. Looking forward teaming up with you.:)</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 427180,
      "author_name": "Mario Parreño Lara",
      "author_url": "",
      "post_date": "2018-11-24T20:13:29.657000",
      "content": "<p>I want a team tooo :)</p>",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "427057": "Hi Kagglers,\n\nI would like to join a team for this competition. As team deadline is close, it's really urgent for me to ask this to you. 😃\n\nWhat I have so far:\n\n - A single CNN model trained on 50k/class, no TTA, no ensemble and giving LB 928. So could be potentially improved.\n\n - A recent GPU\n\n - I am bad at drawing doodles\n\nI would also be grateful to learn from more Kaggle experienced teammates. 🙏\n\nLooking forward teaming up with you.\n\n***\nUpdate with full dataset : LB 935\n",
    "427870": "Hi  Bouteillon,\nI want to join your team and be grateful to learn from your experience. :)\nWhat I have so far:\n 1. Two CNN models of inceptionv3 and resent50 trained on 30k/class, no TTA, no ensemble and giving LB 935 and 932 respectively.\n 2. Two 1080Ti GPU\n 3. Trying to build the right RNN network. At this stage, there is only one LSTM solution that is LB 800 on 30k/class, no TTA, no ensemble \n\nI have made a merge request. Looking forward teaming up with you.:)",
    "427180": "I want a team tooo :)"
  }
}