{
  "id": 116077,
  "title": "Sample using Keras and tf.data.Dataset + visualizations",
  "url": "/competitions/Kannada-MNIST/discussion/116077",
  "author_name": "Martin Görner",
  "post_date": "2019-11-07T00:45:21.758000",
  "votes": 2,
  "comment_count": 0,
  "views": 0,
  "content": "<p>I just posted a sample using Keras and tf.data.Dataset. It has a couple of neat visualizations too, especially remaining mis-classified digits. Enjoy:\n<a href=\"https://www.kaggle.com/martingorner/kannada-martin-keras-5-layer-cnn-with-batch-norm/\">kannada-martin-keras-5-layer-cnn-with-batch-norm</a></p>\n\n<p>(top accuracy is not the goal in this sample)</p>",
  "messages": [
    {
      "id": 667215,
      "postDate": "2019-11-07T00:45:21.760Z",
      "content": "<p>I just posted a sample using Keras and tf.data.Dataset. It has a couple of neat visualizations too, especially remaining mis-classified digits. Enjoy:\n<a href=\"https://www.kaggle.com/martingorner/kannada-martin-keras-5-layer-cnn-with-batch-norm/\">kannada-martin-keras-5-layer-cnn-with-batch-norm</a></p>\n\n<p>(top accuracy is not the goal in this sample)</p>",
      "rawMarkdown": "I just posted a sample using Keras and tf.data.Dataset. It has a couple of neat visualizations too, especially remaining mis-classified digits. Enjoy:\n[kannada-martin-keras-5-layer-cnn-with-batch-norm](https://www.kaggle.com/martingorner/kannada-martin-keras-5-layer-cnn-with-batch-norm/)\n\n(top accuracy is not the goal in this sample)",
      "votes": 2
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "667215": "I just posted a sample using Keras and tf.data.Dataset. It has a couple of neat visualizations too, especially remaining mis-classified digits. Enjoy:\n[kannada-martin-keras-5-layer-cnn-with-batch-norm](https://www.kaggle.com/martingorner/kannada-martin-keras-5-layer-cnn-with-batch-norm/)\n\n(top accuracy is not the goal in this sample)"
  }
}