{
  "id": 155705,
  "title": "power of external data",
  "url": "/competitions/siim-isic-melanoma-classification/discussion/155705",
  "author_name": "",
  "post_date": "2020-06-02T16:58:29.307142Z",
  "votes": 31,
  "comment_count": 8,
  "views": 0,
  "content": "<p>as attached\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F113660%2Fbdc603ef434ad366f150b884ec66bf14%2FSelection_074.png?generation=1591117103480641&amp;alt=media\" alt=\"\">\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F113660%2F2a3dc62fb45240c1907d2995989a1a88%2FSelection_073.png?generation=1591117103533884&amp;alt=media\" alt=\"\"></p>",
  "messages": [
    {
      "id": "871828",
      "postDate": "06/02/2020 16:58:29",
      "content": "<p>as attached\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F113660%2Fbdc603ef434ad366f150b884ec66bf14%2FSelection_074.png?generation=1591117103480641&amp;alt=media\" alt=\"\">\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F113660%2F2a3dc62fb45240c1907d2995989a1a88%2FSelection_073.png?generation=1591117103533884&amp;alt=media\" alt=\"\"></p>",
      "rawMarkdown": "as attached\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F113660%2Fbdc603ef434ad366f150b884ec66bf14%2FSelection_074.png?generation=1591117103480641&amp;alt=media)\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F113660%2F2a3dc62fb45240c1907d2995989a1a88%2FSelection_073.png?generation=1591117103533884&amp;alt=media)",
      "votes": null
    },
    {
      "id": "871869",
      "postDate": "06/02/2020 17:27:42",
      "content": "<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F113660%2F917a4107ed334213a83088bf85152321%2FSelection_077.png?generation=1591118859269263&amp;alt=media\" alt=\"\"></p>",
      "rawMarkdown": "![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F113660%2F917a4107ed334213a83088bf85152321%2FSelection_077.png?generation=1591118859269263&amp;alt=media)",
      "votes": null
    },
    {
      "id": "871972",
      "postDate": "06/02/2020 19:14:20",
      "content": "<p><a href=\"/hengck23\">@hengck23</a> waiting for you to post the data you are using in the external data thread (<a href=\"https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/154296\">https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/154296</a>).</p>\n\n<p>By the way, while our training data has only 584 malignant out of total 33126, the external data at <a href=\"https://www.isic-archive.com/#!/topWithHeader/onlyHeaderTop/gallery\">https://www.isic-archive.com/#!/topWithHeader/onlyHeaderTop/gallery</a> has 2285+ malignant examples, which should help reduce the class imbalance problem.</p>",
      "rawMarkdown": "hengck23 waiting for you to post the data you are using in the external data thread (https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/154296).\n\nBy the way, while our training data has only 584 malignant out of total 33126, the external data at https://www.isic-archive.com/#!/topWithHeader/onlyHeaderTop/gallery has 2285+ malignant examples, which should help reduce the class imbalance problem.",
      "votes": null
    },
    {
      "id": "871994",
      "postDate": "06/02/2020 19:43:00",
      "content": "<p><a href=\"/hengck23\">@hengck23</a> love u</p>",
      "rawMarkdown": "hengck23 love u",
      "votes": null
    },
    {
      "id": "872344",
      "postDate": "06/03/2020 05:32:41",
      "content": "<p>How can i extract this extenral data to our kaggle kernel??</p>",
      "rawMarkdown": "How can i extract this extenral data to our kaggle kernel??",
      "votes": null
    },
    {
      "id": "876299",
      "postDate": "06/06/2020 15:57:36",
      "content": "<p>Hi. Hope you all are fine.\nJust wanted to know your resources like how you train models in such a good variety. \nThanks for the good work btw.</p>",
      "rawMarkdown": "Hi. Hope you all are fine.\nJust wanted to know your resources like how you train models in such a good variety. \nThanks for the good work btw.",
      "votes": null
    },
    {
      "id": "876730",
      "postDate": "06/07/2020 01:22:32",
      "content": "<p>list of all external data:</p>\n\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F113660%2F9f80e566cd8c7d7a960b870461a229ac%2FSelection_106.png?generation=1591497553390727&amp;alt=media\" alt=\"\"></p>",
      "rawMarkdown": "list of all external data:\n\n ![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F113660%2F9f80e566cd8c7d7a960b870461a229ac%2FSelection_106.png?generation=1591497553390727&amp;alt=media)",
      "votes": null
    },
    {
      "id": "876746",
      "postDate": "06/07/2020 02:19:25",
      "content": "<p>I posted TFRecords Kaggle dataset that includes external data <a href=\"https://www.kaggle.com/cdeotte/512x512-melanoma-tfrecords-70k-images\">here</a> and Alex posted JPEG Kaggle dataset that includes external data <a href=\"https://www.kaggle.com/shonenkov/melanoma-merged-external-data-512x512-jpeg\">here</a>. I confirm that these external images have helped my model ensemble.</p>",
      "rawMarkdown": "I posted TFRecords Kaggle dataset that includes external data [here][1] and Alex posted JPEG Kaggle dataset that includes external data [here][2]. I confirm that these external images have helped my model ensemble.\n\n[1]: https://www.kaggle.com/cdeotte/512x512-melanoma-tfrecords-70k-images\n[2]: https://www.kaggle.com/shonenkov/melanoma-merged-external-data-512x512-jpeg",
      "votes": null
    },
    {
      "id": "952325",
      "postDate": "07/30/2020 20:17:36",
      "content": "<p>I have extracted this external data from the ISIC-archive and posted a Kaggle dataset <a href=\"https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/169139\">here</a></p>",
      "rawMarkdown": "I have extracted this external data from the ISIC-archive and posted a Kaggle dataset [here][1]\n\n[1]: https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/169139",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 871869,
      "author_name": "hengck23",
      "author_url": "",
      "post_date": "06/02/2020 17:27:42",
      "content": "<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F113660%2F917a4107ed334213a83088bf85152321%2FSelection_077.png?generation=1591118859269263&amp;alt=media\" alt=\"\"></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 871972,
      "author_name": "sirishks",
      "author_url": "",
      "post_date": "06/02/2020 19:14:20",
      "content": "<p><a href=\"/hengck23\">@hengck23</a> waiting for you to post the data you are using in the external data thread (<a href=\"https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/154296\">https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/154296</a>).</p>\n\n<p>By the way, while our training data has only 584 malignant out of total 33126, the external data at <a href=\"https://www.isic-archive.com/#!/topWithHeader/onlyHeaderTop/gallery\">https://www.isic-archive.com/#!/topWithHeader/onlyHeaderTop/gallery</a> has 2285+ malignant examples, which should help reduce the class imbalance problem.</p>",
      "votes": null,
      "replies": [
        {
          "id": 872344,
          "author_name": "msharuk589",
          "author_url": "",
          "post_date": "06/03/2020 05:32:41",
          "content": "<p>How can i extract this extenral data to our kaggle kernel??</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 952325,
          "author_name": "cdeotte",
          "author_url": "",
          "post_date": "07/30/2020 20:17:36",
          "content": "<p>I have extracted this external data from the ISIC-archive and posted a Kaggle dataset <a href=\"https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/169139\">here</a></p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 871994,
      "author_name": "romanweilguny",
      "author_url": "",
      "post_date": "06/02/2020 19:43:00",
      "content": "<p><a href=\"/hengck23\">@hengck23</a> love u</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 876299,
      "author_name": "micheomaano",
      "author_url": "",
      "post_date": "06/06/2020 15:57:36",
      "content": "<p>Hi. Hope you all are fine.\nJust wanted to know your resources like how you train models in such a good variety. \nThanks for the good work btw.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 876730,
      "author_name": "hengck23",
      "author_url": "",
      "post_date": "06/07/2020 01:22:32",
      "content": "<p>list of all external data:</p>\n\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F113660%2F9f80e566cd8c7d7a960b870461a229ac%2FSelection_106.png?generation=1591497553390727&amp;alt=media\" alt=\"\"></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 876746,
      "author_name": "cdeotte",
      "author_url": "",
      "post_date": "06/07/2020 02:19:25",
      "content": "<p>I posted TFRecords Kaggle dataset that includes external data <a href=\"https://www.kaggle.com/cdeotte/512x512-melanoma-tfrecords-70k-images\">here</a> and Alex posted JPEG Kaggle dataset that includes external data <a href=\"https://www.kaggle.com/shonenkov/melanoma-merged-external-data-512x512-jpeg\">here</a>. I confirm that these external images have helped my model ensemble.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "871828": "as attached\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F113660%2Fbdc603ef434ad366f150b884ec66bf14%2FSelection_074.png?generation=1591117103480641&amp;alt=media)\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F113660%2F2a3dc62fb45240c1907d2995989a1a88%2FSelection_073.png?generation=1591117103533884&amp;alt=media)",
    "871869": "![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F113660%2F917a4107ed334213a83088bf85152321%2FSelection_077.png?generation=1591118859269263&amp;alt=media)",
    "871972": "hengck23 waiting for you to post the data you are using in the external data thread (https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/154296).\n\nBy the way, while our training data has only 584 malignant out of total 33126, the external data at https://www.isic-archive.com/#!/topWithHeader/onlyHeaderTop/gallery has 2285+ malignant examples, which should help reduce the class imbalance problem.",
    "871994": "hengck23 love u",
    "872344": "How can i extract this extenral data to our kaggle kernel??",
    "876299": "Hi. Hope you all are fine.\nJust wanted to know your resources like how you train models in such a good variety. \nThanks for the good work btw.",
    "876730": "list of all external data:\n\n ![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F113660%2F9f80e566cd8c7d7a960b870461a229ac%2FSelection_106.png?generation=1591497553390727&amp;alt=media)",
    "876746": "I posted TFRecords Kaggle dataset that includes external data [here][1] and Alex posted JPEG Kaggle dataset that includes external data [here][2]. I confirm that these external images have helped my model ensemble.\n\n[1]: https://www.kaggle.com/cdeotte/512x512-melanoma-tfrecords-70k-images\n[2]: https://www.kaggle.com/shonenkov/melanoma-merged-external-data-512x512-jpeg",
    "952325": "I have extracted this external data from the ISIC-archive and posted a Kaggle dataset [here][1]\n\n[1]: https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/169139"
  },
  "source": "meta"
}