{
  "id": 227445,
  "title": "Silver with a lots of learning!",
  "url": "/competitions/ranzcr-clip-catheter-line-classification/discussion/227445",
  "author_name": "Divyansh Agrawal",
  "post_date": "2021-03-20T13:42:54.625000",
  "votes": 7,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Now that the competition is over, I'd like to extend a bow of gratitude towards my humble and hard working team mates <a href=\"https://www.kaggle.com/frtgnn\" target=\"_blank\">@frtgnn</a>, <a href=\"https://www.kaggle.com/datafan07\" target=\"_blank\">@datafan07</a>, <a href=\"https://www.kaggle.com/pukkinming\" target=\"_blank\">@pukkinming</a>. You guys have been really awesome throughout the competition, plus for your patience! On the way, I learnt how powerful could be Resnet200d models and got to try hands-on NFNets, yet another new player in image transfer learning! Although the training was tough and time taking, it was worth it, eventually landed up with the silver. Also the image size for training was high (640) so it was the courtesy of Z by HP, for providing us with the beast to even try 2048 image sizes!! Perhaps it was another team member for us 😅 I was reluctant to use TPUs before, got working with it in this competition. <br>\nP.S. The Kaggle GPU quota sucks up real fast! 😂 And lastly really sorry to my boys if I did something along the way you didn't like!🙏 Hope to be working with you guys again soon.✌️</p>",
  "messages": [
    {
      "id": 1246165,
      "postDate": "2021-03-20T13:42:54.627Z",
      "content": "<p>Now that the competition is over, I'd like to extend a bow of gratitude towards my humble and hard working team mates <a href=\"https://www.kaggle.com/frtgnn\" target=\"_blank\">@frtgnn</a>, <a href=\"https://www.kaggle.com/datafan07\" target=\"_blank\">@datafan07</a>, <a href=\"https://www.kaggle.com/pukkinming\" target=\"_blank\">@pukkinming</a>. You guys have been really awesome throughout the competition, plus for your patience! On the way, I learnt how powerful could be Resnet200d models and got to try hands-on NFNets, yet another new player in image transfer learning! Although the training was tough and time taking, it was worth it, eventually landed up with the silver. Also the image size for training was high (640) so it was the courtesy of Z by HP, for providing us with the beast to even try 2048 image sizes!! Perhaps it was another team member for us 😅 I was reluctant to use TPUs before, got working with it in this competition. <br>\nP.S. The Kaggle GPU quota sucks up real fast! 😂 And lastly really sorry to my boys if I did something along the way you didn't like!🙏 Hope to be working with you guys again soon.✌️</p>",
      "rawMarkdown": "Now that the competition is over, I'd like to extend a bow of gratitude towards my humble and hard working team mates @frtgnn, @datafan07, @pukkinming. You guys have been really awesome throughout the competition, plus for your patience! On the way, I learnt how powerful could be Resnet200d models and got to try hands-on NFNets, yet another new player in image transfer learning! Although the training was tough and time taking, it was worth it, eventually landed up with the silver. Also the image size for training was high (640) so it was the courtesy of Z by HP, for providing us with the beast to even try 2048 image sizes!! Perhaps it was another team member for us 😅 I was reluctant to use TPUs before, got working with it in this competition. \nP.S. The Kaggle GPU quota sucks up real fast! 😂 And lastly really sorry to my boys if I did something along the way you didn't like!🙏 Hope to be working with you guys again soon.✌️",
      "votes": 7
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1246165": "Now that the competition is over, I'd like to extend a bow of gratitude towards my humble and hard working team mates @frtgnn, @datafan07, @pukkinming. You guys have been really awesome throughout the competition, plus for your patience! On the way, I learnt how powerful could be Resnet200d models and got to try hands-on NFNets, yet another new player in image transfer learning! Although the training was tough and time taking, it was worth it, eventually landed up with the silver. Also the image size for training was high (640) so it was the courtesy of Z by HP, for providing us with the beast to even try 2048 image sizes!! Perhaps it was another team member for us 😅 I was reluctant to use TPUs before, got working with it in this competition. \nP.S. The Kaggle GPU quota sucks up real fast! 😂 And lastly really sorry to my boys if I did something along the way you didn't like!🙏 Hope to be working with you guys again soon.✌️"
  }
}