{
  "id": 246209,
  "title": "为啥交上来的分数超级低",
  "url": "/competitions/classify-leaves/discussion/246209",
  "author_name": "",
  "post_date": "2021-06-14T13:29:20.769831100Z",
  "votes": 5,
  "comment_count": 5,
  "views": 0,
  "content": "<p>在本地训练显示验证集的acc也不是很低，可以测试集交上来就0.00几</p>",
  "messages": [
    {
      "id": "1349036",
      "postDate": "06/14/2021 13:29:20",
      "content": "<p>在本地训练显示验证集的acc也不是很低，可以测试集交上来就0.00几</p>",
      "rawMarkdown": "在本地训练显示验证集的acc也不是很低，可以测试集交上来就0.00几",
      "votes": null
    },
    {
      "id": "1349052",
      "postDate": "06/14/2021 13:40:27",
      "content": "<p>检查你提交的内容，是否乱序或者image与label对不上</p>",
      "rawMarkdown": "检查你提交的内容，是否乱序或者image与label对不上",
      "votes": null
    },
    {
      "id": "1547557",
      "postDate": "10/17/2021 10:48:04",
      "content": "<p>请问这个问题解决了吗😳，怎么样啊</p>",
      "rawMarkdown": "请问这个问题解决了吗😳，怎么样啊",
      "votes": null
    },
    {
      "id": "1547589",
      "postDate": "10/17/2021 11:41:50",
      "content": "<p>家人们真相了，原来是dataloader shuffle=True了</p>",
      "rawMarkdown": "家人们真相了，原来是dataloader shuffle=True了",
      "votes": null
    },
    {
      "id": "1694591",
      "postDate": "02/17/2022 15:00:35",
      "content": "<p>！！看到你这个我就知道我那错了。。</p>",
      "rawMarkdown": "！！看到你这个我就知道我那错了。。",
      "votes": null
    },
    {
      "id": "1854805",
      "postDate": "07/14/2022 03:25:11",
      "content": "<p>我也遇到了类似的问题，使用预训练的resnet模型，本地训练和验证都0.95+， 但是提交上来只有不到0.3，后面发现是测试数据集没有做normalization !!!</p>",
      "rawMarkdown": "我也遇到了类似的问题，使用预训练的resnet模型，本地训练和验证都0.95+， 但是提交上来只有不到0.3，后面发现是测试数据集没有做normalization !!!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1349052,
      "author_name": "byeblack",
      "author_url": "",
      "post_date": "06/14/2021 13:40:27",
      "content": "<p>检查你提交的内容，是否乱序或者image与label对不上</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1547557,
      "author_name": "jbinary",
      "author_url": "",
      "post_date": "10/17/2021 10:48:04",
      "content": "<p>请问这个问题解决了吗😳，怎么样啊</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1547589,
      "author_name": "jbinary",
      "author_url": "",
      "post_date": "10/17/2021 11:41:50",
      "content": "<p>家人们真相了，原来是dataloader shuffle=True了</p>",
      "votes": null,
      "replies": [
        {
          "id": 1694591,
          "author_name": "gavinhechu",
          "author_url": "",
          "post_date": "02/17/2022 15:00:35",
          "content": "<p>！！看到你这个我就知道我那错了。。</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1854805,
      "author_name": "staneyffer",
      "author_url": "",
      "post_date": "07/14/2022 03:25:11",
      "content": "<p>我也遇到了类似的问题，使用预训练的resnet模型，本地训练和验证都0.95+， 但是提交上来只有不到0.3，后面发现是测试数据集没有做normalization !!!</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1349036": "在本地训练显示验证集的acc也不是很低，可以测试集交上来就0.00几",
    "1349052": "检查你提交的内容，是否乱序或者image与label对不上",
    "1547557": "请问这个问题解决了吗😳，怎么样啊",
    "1547589": "家人们真相了，原来是dataloader shuffle=True了",
    "1694591": "！！看到你这个我就知道我那错了。。",
    "1854805": "我也遇到了类似的问题，使用预训练的resnet模型，本地训练和验证都0.95+， 但是提交上来只有不到0.3，后面发现是测试数据集没有做normalization !!!"
  },
  "source": "meta"
}