{
  "id": 295630,
  "title": "TensorFlow OD API Pipeline [0.474]",
  "url": "/competitions/tensorflow-great-barrier-reef/discussion/295630",
  "author_name": "Ravi Shah",
  "post_date": "2021-12-17T01:28:15.669000",
  "votes": 3,
  "comment_count": 0,
  "views": 0,
  "content": "<p>I’ve recently published my 0.474 training and inference pipeline using the TensorFlow Object Detection API. This is not the highest scoring public kernel, but it is eligible for the TensorFlow Performance Prize. I use a Faster-RCNN with a Resnet-101 feature extractor. I applied a few augmentations, used a momentum optimizer, and split with a group k-fold on sequence. </p>\n<p><a href=\"https://www.kaggle.com/ravishah1/cots-faster-rcnn-training-w-tf-2-0-od-api-0-474\" target=\"_blank\">Here is the training notebook</a><br>\n<a href=\"https://www.kaggle.com/ravishah1/cots-faster-rcnn-inference-tf-2-0-od-api-0-474\" target=\"_blank\">Here is the inference notebook</a><br>\nFor more details on how to improve these types of model, <a href=\"https://www.kaggle.com/c/tensorflow-great-barrier-reef/discussion/293305\" target=\"_blank\">read this to better understand configs</a></p>",
  "messages": [
    {
      "id": 1620627,
      "postDate": "2021-12-17T01:28:15.670Z",
      "content": "<p>I’ve recently published my 0.474 training and inference pipeline using the TensorFlow Object Detection API. This is not the highest scoring public kernel, but it is eligible for the TensorFlow Performance Prize. I use a Faster-RCNN with a Resnet-101 feature extractor. I applied a few augmentations, used a momentum optimizer, and split with a group k-fold on sequence. </p>\n<p><a href=\"https://www.kaggle.com/ravishah1/cots-faster-rcnn-training-w-tf-2-0-od-api-0-474\" target=\"_blank\">Here is the training notebook</a><br>\n<a href=\"https://www.kaggle.com/ravishah1/cots-faster-rcnn-inference-tf-2-0-od-api-0-474\" target=\"_blank\">Here is the inference notebook</a><br>\nFor more details on how to improve these types of model, <a href=\"https://www.kaggle.com/c/tensorflow-great-barrier-reef/discussion/293305\" target=\"_blank\">read this to better understand configs</a></p>",
      "rawMarkdown": "I’ve recently published my 0.474 training and inference pipeline using the TensorFlow Object Detection API. This is not the highest scoring public kernel, but it is eligible for the TensorFlow Performance Prize. I use a Faster-RCNN with a Resnet-101 feature extractor. I applied a few augmentations, used a momentum optimizer, and split with a group k-fold on sequence. \n\n[Here is the training notebook](https://www.kaggle.com/ravishah1/cots-faster-rcnn-training-w-tf-2-0-od-api-0-474)\n[Here is the inference notebook](https://www.kaggle.com/ravishah1/cots-faster-rcnn-inference-tf-2-0-od-api-0-474)\nFor more details on how to improve these types of model, [read this to better understand configs](https://www.kaggle.com/c/tensorflow-great-barrier-reef/discussion/293305)",
      "votes": 3
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1620627": "I’ve recently published my 0.474 training and inference pipeline using the TensorFlow Object Detection API. This is not the highest scoring public kernel, but it is eligible for the TensorFlow Performance Prize. I use a Faster-RCNN with a Resnet-101 feature extractor. I applied a few augmentations, used a momentum optimizer, and split with a group k-fold on sequence. \n\n[Here is the training notebook](https://www.kaggle.com/ravishah1/cots-faster-rcnn-training-w-tf-2-0-od-api-0-474)\n[Here is the inference notebook](https://www.kaggle.com/ravishah1/cots-faster-rcnn-inference-tf-2-0-od-api-0-474)\nFor more details on how to improve these types of model, [read this to better understand configs](https://www.kaggle.com/c/tensorflow-great-barrier-reef/discussion/293305)"
  }
}