{
  "id": 166126,
  "title": "Data augmentation EDA",
  "url": "/competitions/siim-isic-melanoma-classification/discussion/166126",
  "author_name": "DimitreOliveira",
  "post_date": "2020-07-11T21:26:26.531000",
  "votes": 8,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Hey everyone, I was struggling a little at choosing augmentations that make sense to the problem domain and specially tunning its hyperparameters, so I created a <a href=\"https://www.kaggle.com/dimitreoliveira/melanoma-classification-augmentations-eda\">very basic notebook</a> to explore and analyze the impact os different kinds of augmentations at the dataset images, it is very modular and easy to customize, this is one sample image that uses many augmentations:</p>\n\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F1182060%2F27585cfe731dc2615c4d30b80b9e443f%2FScreenshot%20from%202020-07-12%2000-06-01.png?generation=1594523188386032&amp;alt=media\" alt=\"\"></p>",
  "messages": [
    {
      "id": 925170,
      "postDate": "2020-07-11T21:26:26.533Z",
      "content": "<p>Hey everyone, I was struggling a little at choosing augmentations that make sense to the problem domain and specially tunning its hyperparameters, so I created a <a href=\"https://www.kaggle.com/dimitreoliveira/melanoma-classification-augmentations-eda\">very basic notebook</a> to explore and analyze the impact os different kinds of augmentations at the dataset images, it is very modular and easy to customize, this is one sample image that uses many augmentations:</p>\n\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F1182060%2F27585cfe731dc2615c4d30b80b9e443f%2FScreenshot%20from%202020-07-12%2000-06-01.png?generation=1594523188386032&amp;alt=media\" alt=\"\"></p>",
      "rawMarkdown": "Hey everyone, I was struggling a little at choosing augmentations that make sense to the problem domain and specially tunning its hyperparameters, so I created a [very basic notebook](https://www.kaggle.com/dimitreoliveira/melanoma-classification-augmentations-eda) to explore and analyze the impact os different kinds of augmentations at the dataset images, it is very modular and easy to customize, this is one sample image that uses many augmentations:\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F1182060%2F27585cfe731dc2615c4d30b80b9e443f%2FScreenshot%20from%202020-07-12%2000-06-01.png?generation=1594523188386032&amp;alt=media)",
      "votes": 7
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "925170": "Hey everyone, I was struggling a little at choosing augmentations that make sense to the problem domain and specially tunning its hyperparameters, so I created a [very basic notebook](https://www.kaggle.com/dimitreoliveira/melanoma-classification-augmentations-eda) to explore and analyze the impact os different kinds of augmentations at the dataset images, it is very modular and easy to customize, this is one sample image that uses many augmentations:\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F1182060%2F27585cfe731dc2615c4d30b80b9e443f%2FScreenshot%20from%202020-07-12%2000-06-01.png?generation=1594523188386032&amp;alt=media)"
  }
}