{
  "id": 449198,
  "title": "Are you using any pretrained model?",
  "url": "/competitions/UBC-OCEAN/discussion/449198",
  "author_name": "Tushar",
  "post_date": "2023-10-23T13:42:41.720000",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Are there any pretrained model, that gives satisfactory performance on this dataset? </p>",
  "messages": [
    {
      "id": 2508551,
      "postDate": "2023-11-01T19:31:04.180Z",
      "content": "<p>You may start with any standard model and pull its weight for ImageNet, as you always need to train it on this dataset…<br>\nI have been using TIMM models, but you can use any other - <a href=\"https://kaggle.com/code/jirkaborovec/cancer-subtype-tiles-w-lightning-timm-models\" target=\"_blank\">https://kaggle.com/code/jirkaborovec/cancer-subtype-tiles-w-lightning-timm-models</a></p>",
      "rawMarkdown": "You may start with any standard model and pull its weight for ImageNet, as you always need to train it on this dataset...\nI have been using TIMM models, but you can use any other - https://kaggle.com/code/jirkaborovec/cancer-subtype-tiles-w-lightning-timm-models",
      "votes": 1
    },
    {
      "id": 2493581,
      "postDate": "2023-10-23T13:42:41.720Z",
      "content": "<p>Are there any pretrained model, that gives satisfactory performance on this dataset? </p>",
      "rawMarkdown": "Are there any pretrained model, that gives satisfactory performance on this dataset? ",
      "votes": 1
    }
  ],
  "comments": [
    {
      "id": 2508551,
      "author_name": "Jirka",
      "author_url": "",
      "post_date": "2023-11-01T19:31:04.180000",
      "content": "<p>You may start with any standard model and pull its weight for ImageNet, as you always need to train it on this dataset…<br>\nI have been using TIMM models, but you can use any other - <a href=\"https://kaggle.com/code/jirkaborovec/cancer-subtype-tiles-w-lightning-timm-models\" target=\"_blank\">https://kaggle.com/code/jirkaborovec/cancer-subtype-tiles-w-lightning-timm-models</a></p>",
      "votes": 1,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2508551": "You may start with any standard model and pull its weight for ImageNet, as you always need to train it on this dataset...\nI have been using TIMM models, but you can use any other - https://kaggle.com/code/jirkaborovec/cancer-subtype-tiles-w-lightning-timm-models",
    "2493581": "Are there any pretrained model, that gives satisfactory performance on this dataset? "
  }
}