{
  "id": 174453,
  "title": "Numerous Output files ",
  "url": "/competitions/siim-isic-melanoma-classification/discussion/174453",
  "author_name": "MhdSharuk",
  "post_date": "2020-08-13T15:19:45.267000",
  "votes": 0,
  "comment_count": 0,
  "views": 0,
  "content": "<p>This week i have trained on (192,256,384,512,768) image sizes on each of the following efficient net b[3,4,5,6,7] without meta data and Tree based learning algorithms of meta data….Since there is a lots of submission files…how do i need to approach the ensembling method on these files…?<br>\n(Finding good weights on like approximately for almost 30 submission file is not the right way like <a href=\"https://www.kaggle.com/sirishks\" target=\"_blank\">@sirishks</a> mentioned in his one ensembling method posts)<br>\nHope others will also get knowledge from this…</p>",
  "messages": [
    {
      "id": 969263,
      "postDate": "2020-08-13T15:19:45.267Z",
      "content": "<p>This week i have trained on (192,256,384,512,768) image sizes on each of the following efficient net b[3,4,5,6,7] without meta data and Tree based learning algorithms of meta data….Since there is a lots of submission files…how do i need to approach the ensembling method on these files…?<br>\n(Finding good weights on like approximately for almost 30 submission file is not the right way like <a href=\"https://www.kaggle.com/sirishks\" target=\"_blank\">@sirishks</a> mentioned in his one ensembling method posts)<br>\nHope others will also get knowledge from this…</p>",
      "rawMarkdown": "This week i have trained on (192,256,384,512,768) image sizes on each of the following efficient net b[3,4,5,6,7] without meta data and Tree based learning algorithms of meta data....Since there is a lots of submission files...how do i need to approach the ensembling method on these files...?\n(Finding good weights on like approximately for almost 30 submission file is not the right way like @sirishks mentioned in his one ensembling method posts)\nHope others will also get knowledge from this..."
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "969263": "This week i have trained on (192,256,384,512,768) image sizes on each of the following efficient net b[3,4,5,6,7] without meta data and Tree based learning algorithms of meta data....Since there is a lots of submission files...how do i need to approach the ensembling method on these files...?\n(Finding good weights on like approximately for almost 30 submission file is not the right way like @sirishks mentioned in his one ensembling method posts)\nHope others will also get knowledge from this..."
  }
}