{
  "id": 228329,
  "title": "Which encoder performs well?",
  "url": "/competitions/hubmap-kidney-segmentation/discussion/228329",
  "author_name": "",
  "post_date": "2021-03-24T07:28:44.411951800Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I tried resnet and efficientnet and found that efficientnet is better this time.<br>\nAnd also I'm curious why resnet and efficientnet are so popular on kaggle (if it's true), there are way more models to choose.</p>",
  "messages": [
    {
      "id": "1250653",
      "postDate": "03/24/2021 07:28:44",
      "content": "<p>I tried resnet and efficientnet and found that efficientnet is better this time.<br>\nAnd also I'm curious why resnet and efficientnet are so popular on kaggle (if it's true), there are way more models to choose.</p>",
      "rawMarkdown": "I tried resnet and efficientnet and found that efficientnet is better this time.\nAnd also I'm curious why resnet and efficientnet are so popular on kaggle (if it's true), there are way more models to choose.",
      "votes": null
    },
    {
      "id": "1250677",
      "postDate": "03/24/2021 07:43:35",
      "content": "<p>I would recommend you read the EfficientNet paper to learn why they are so prevalent on Kaggle: <a href=\"https://arxiv.org/abs/1905.11946\" target=\"_blank\">https://arxiv.org/abs/1905.11946</a></p>\n<p>As for which encoder is best suited for this task, it may not be an EfficientNet, as in recent competitions different variants of ResNets have worked well or even better for medical imaging tasks</p>",
      "rawMarkdown": "I would recommend you read the EfficientNet paper to learn why they are so prevalent on Kaggle: https://arxiv.org/abs/1905.11946\n\nAs for which encoder is best suited for this task, it may not be an EfficientNet, as in recent competitions different variants of ResNets have worked well or even better for medical imaging tasks",
      "votes": null
    },
    {
      "id": "1250697",
      "postDate": "03/24/2021 08:03:02",
      "content": "<p>Thank you for your information, I will try different variants of resnet later.</p>",
      "rawMarkdown": "Thank you for your information, I will try different variants of resnet later.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1250677,
      "author_name": "anjum48",
      "author_url": "",
      "post_date": "03/24/2021 07:43:35",
      "content": "<p>I would recommend you read the EfficientNet paper to learn why they are so prevalent on Kaggle: <a href=\"https://arxiv.org/abs/1905.11946\" target=\"_blank\">https://arxiv.org/abs/1905.11946</a></p>\n<p>As for which encoder is best suited for this task, it may not be an EfficientNet, as in recent competitions different variants of ResNets have worked well or even better for medical imaging tasks</p>",
      "votes": null,
      "replies": [
        {
          "id": 1250697,
          "author_name": "dylanliuofficial",
          "author_url": "",
          "post_date": "03/24/2021 08:03:02",
          "content": "<p>Thank you for your information, I will try different variants of resnet later.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1250653": "I tried resnet and efficientnet and found that efficientnet is better this time.\nAnd also I'm curious why resnet and efficientnet are so popular on kaggle (if it's true), there are way more models to choose.",
    "1250677": "I would recommend you read the EfficientNet paper to learn why they are so prevalent on Kaggle: https://arxiv.org/abs/1905.11946\n\nAs for which encoder is best suited for this task, it may not be an EfficientNet, as in recent competitions different variants of ResNets have worked well or even better for medical imaging tasks",
    "1250697": "Thank you for your information, I will try different variants of resnet later."
  },
  "source": "meta"
}