{
  "id": 200949,
  "title": "TF with TPU EfficientUNetB7",
  "url": "/competitions/hubmap-kidney-segmentation/discussion/200949",
  "author_name": "Wojtek Rosa",
  "post_date": "2020-12-02T13:40:09.978000",
  "votes": 3,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Hello<br>\nI just published a notebook <a href=\"https://www.kaggle.com/wrrosa/hubmap-tf-with-tpu-efficientunetb7-512x512-train\" target=\"_blank\">here</a> containing training pipeline for tensorflow with tpu efficientunetB7  (many more backbones from the qubvel package segmentation_models are available). The obtained weights and model can be very useful for inference. <br>\nGreetings!</p>",
  "messages": [
    {
      "id": 1099626,
      "postDate": "2020-12-02T13:40:09.980Z",
      "content": "<p>Hello<br>\nI just published a notebook <a href=\"https://www.kaggle.com/wrrosa/hubmap-tf-with-tpu-efficientunetb7-512x512-train\" target=\"_blank\">here</a> containing training pipeline for tensorflow with tpu efficientunetB7  (many more backbones from the qubvel package segmentation_models are available). The obtained weights and model can be very useful for inference. <br>\nGreetings!</p>",
      "rawMarkdown": "Hello\nI just published a notebook [here](https://www.kaggle.com/wrrosa/hubmap-tf-with-tpu-efficientunetb7-512x512-train) containing training pipeline for tensorflow with tpu efficientunetB7  (many more backbones from the qubvel package segmentation_models are available). The obtained weights and model can be very useful for inference. \nGreetings!",
      "votes": 3
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1099626": "Hello\nI just published a notebook [here](https://www.kaggle.com/wrrosa/hubmap-tf-with-tpu-efficientunetb7-512x512-train) containing training pipeline for tensorflow with tpu efficientunetB7  (many more backbones from the qubvel package segmentation_models are available). The obtained weights and model can be very useful for inference. \nGreetings!"
  }
}