{
  "id": 307623,
  "title": "Efficientnet-b0 Stage 2 Post Process",
  "url": "/competitions/tensorflow-great-barrier-reef/discussion/307623",
  "author_name": "Dewei Chen",
  "post_date": "2022-02-15T02:28:49.441000",
  "votes": 12,
  "comment_count": 0,
  "views": 0,
  "content": "<p>I build a model with two-stage efficientnet-b0 for post-process.<br>\nUnfortunately, our team cannot implement it on our model on time.<br>\nIt work very well on local test.</p>\n<p><strong>The basis idea, is from RCNN, which has a stage-2 resnet for classification.\nSame the idea, I built it for yolov5.</strong><br>\nThe notebook can run is here:<br>\n<a href=\"https://www.kaggle.com/dwchen/final-yolov5tracking-2stage/notebook\" target=\"_blank\">https://www.kaggle.com/dwchen/final-yolov5tracking-2stage/notebook</a><br>\nThe train code:<br>\n<a href=\"https://www.kaggle.com/dwchen/tez-starfish-training\" target=\"_blank\">https://www.kaggle.com/dwchen/tez-starfish-training</a><br>\nThe inference test code:<br>\n<a href=\"https://www.kaggle.com/dwchen/tez-starfish-inference\" target=\"_blank\">https://www.kaggle.com/dwchen/tez-starfish-inference</a><br>\nThe original dataset:<br>\n<a href=\"https://www.kaggle.com/c/tensorflow-great-barrier-reef/discussion/305367\" target=\"_blank\">https://www.kaggle.com/c/tensorflow-great-barrier-reef/discussion/305367</a></p>\n<p>Hope it can helped you in future</p>",
  "messages": [
    {
      "id": 1690572,
      "postDate": "2022-02-15T02:28:49.440Z",
      "content": "<p>I build a model with two-stage efficientnet-b0 for post-process.<br>\nUnfortunately, our team cannot implement it on our model on time.<br>\nIt work very well on local test.</p>\n<p><strong>The basis idea, is from RCNN, which has a stage-2 resnet for classification.\nSame the idea, I built it for yolov5.</strong><br>\nThe notebook can run is here:<br>\n<a href=\"https://www.kaggle.com/dwchen/final-yolov5tracking-2stage/notebook\" target=\"_blank\">https://www.kaggle.com/dwchen/final-yolov5tracking-2stage/notebook</a><br>\nThe train code:<br>\n<a href=\"https://www.kaggle.com/dwchen/tez-starfish-training\" target=\"_blank\">https://www.kaggle.com/dwchen/tez-starfish-training</a><br>\nThe inference test code:<br>\n<a href=\"https://www.kaggle.com/dwchen/tez-starfish-inference\" target=\"_blank\">https://www.kaggle.com/dwchen/tez-starfish-inference</a><br>\nThe original dataset:<br>\n<a href=\"https://www.kaggle.com/c/tensorflow-great-barrier-reef/discussion/305367\" target=\"_blank\">https://www.kaggle.com/c/tensorflow-great-barrier-reef/discussion/305367</a></p>\n<p>Hope it can helped you in future</p>",
      "rawMarkdown": "I build a model with two-stage efficientnet-b0 for post-process.\nUnfortunately, our team cannot implement it on our model on time.\nIt work very well on local test.\n\n**The basis idea, is from RCNN, which has a stage-2 resnet for classification.\nSame the idea, I built it for yolov5.**\nThe notebook can run is here:\nhttps://www.kaggle.com/dwchen/final-yolov5tracking-2stage/notebook\nThe train code:\nhttps://www.kaggle.com/dwchen/tez-starfish-training\nThe inference test code:\nhttps://www.kaggle.com/dwchen/tez-starfish-inference\nThe original dataset:\nhttps://www.kaggle.com/c/tensorflow-great-barrier-reef/discussion/305367\n\nHope it can helped you in future",
      "votes": 12
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1690572": "I build a model with two-stage efficientnet-b0 for post-process.\nUnfortunately, our team cannot implement it on our model on time.\nIt work very well on local test.\n\n**The basis idea, is from RCNN, which has a stage-2 resnet for classification.\nSame the idea, I built it for yolov5.**\nThe notebook can run is here:\nhttps://www.kaggle.com/dwchen/final-yolov5tracking-2stage/notebook\nThe train code:\nhttps://www.kaggle.com/dwchen/tez-starfish-training\nThe inference test code:\nhttps://www.kaggle.com/dwchen/tez-starfish-inference\nThe original dataset:\nhttps://www.kaggle.com/c/tensorflow-great-barrier-reef/discussion/305367\n\nHope it can helped you in future"
  }
}