{
  "id": 175341,
  "title": "CV / Private LB score",
  "url": "/competitions/siim-isic-melanoma-classification/discussion/175341",
  "author_name": "",
  "post_date": "2020-08-18T01:00:12.863529600Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Trust CV seemed to be the most important thing in this competition.<br>\nWhat is your CV / <strong>Private</strong> Score?<br>\nMine:<br>\nCV 0.9476 / Private 0.9399 (Public 0.9520)</p>\n<ul>\n<li>Validation: TripleStratifiedKFold(5fold)</li>\n<li>Image_size: 384, 512, 768</li>\n<li>Model: EfficientNet B3, B4, B5, B6</li>\n<li>Loss: focal loss, BCE</li>\n<li>Optimizer: Radam</li>\n<li>2018, 2017 external data</li>\n<li>Stacking --&gt; weighted average with meta data model(LGBM)</li>\n</ul>",
  "messages": [
    {
      "id": "974506",
      "postDate": "08/18/2020 01:00:12",
      "content": "<p>Trust CV seemed to be the most important thing in this competition.<br>\nWhat is your CV / <strong>Private</strong> Score?<br>\nMine:<br>\nCV 0.9476 / Private 0.9399 (Public 0.9520)</p>\n<ul>\n<li>Validation: TripleStratifiedKFold(5fold)</li>\n<li>Image_size: 384, 512, 768</li>\n<li>Model: EfficientNet B3, B4, B5, B6</li>\n<li>Loss: focal loss, BCE</li>\n<li>Optimizer: Radam</li>\n<li>2018, 2017 external data</li>\n<li>Stacking --&gt; weighted average with meta data model(LGBM)</li>\n</ul>",
      "rawMarkdown": "Trust CV seemed to be the most important thing in this competition.\nWhat is your CV / **Private** Score?\nMine:\nCV 0.9476 / Private 0.9399 (Public 0.9520)\n- Validation: TripleStratifiedKFold(5fold)\n- Image_size: 384, 512, 768\n- Model: EfficientNet B3, B4, B5, B6\n- Loss: focal loss, BCE\n- Optimizer: Radam\n- 2018, 2017 external data\n- Stacking --> weighted average with meta data model(LGBM)",
      "votes": null
    },
    {
      "id": "974561",
      "postDate": "08/18/2020 01:31:56",
      "content": "<p><a href=\"https://www.kaggle.com/amanatsu\" target=\"_blank\">@amanatsu</a> my results are here: <br>\n<a href=\"https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/175330\" target=\"_blank\">https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/175330</a></p>",
      "rawMarkdown": "amanatsu my results are here: \nhttps://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/175330",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 974561,
      "author_name": "mpsampat",
      "author_url": "",
      "post_date": "08/18/2020 01:31:56",
      "content": "<p><a href=\"https://www.kaggle.com/amanatsu\" target=\"_blank\">@amanatsu</a> my results are here: <br>\n<a href=\"https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/175330\" target=\"_blank\">https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/175330</a></p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "974506": "Trust CV seemed to be the most important thing in this competition.\nWhat is your CV / **Private** Score?\nMine:\nCV 0.9476 / Private 0.9399 (Public 0.9520)\n- Validation: TripleStratifiedKFold(5fold)\n- Image_size: 384, 512, 768\n- Model: EfficientNet B3, B4, B5, B6\n- Loss: focal loss, BCE\n- Optimizer: Radam\n- 2018, 2017 external data\n- Stacking --> weighted average with meta data model(LGBM)",
    "974561": "amanatsu my results are here: \nhttps://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/175330"
  },
  "source": "meta"
}