{
  "id": 74553,
  "title": "Protein classification (HPA) - Not at all improving accuracy ",
  "url": "/competitions/human-protein-atlas-image-classification/discussion/74553",
  "author_name": "",
  "post_date": "2018-12-13T11:48:27.997852700Z",
  "votes": null,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I tried desnet 169 with the given macro f1 score login. Used 4channels as well 3channels with 512x512 resolution. But, the model's accuracy did not improve. Able to get 0.25 score only. </p>\n\n<p>Any suggestion, how do improve the models performance?</p>",
  "messages": [
    {
      "id": "438269",
      "postDate": "12/13/2018 11:48:27",
      "content": "<p>I tried desnet 169 with the given macro f1 score login. Used 4channels as well 3channels with 512x512 resolution. But, the model's accuracy did not improve. Able to get 0.25 score only. </p>\n\n<p>Any suggestion, how do improve the models performance?</p>",
      "rawMarkdown": "I tried desnet 169 with the given macro f1 score login. Used 4channels as well 3channels with 512x512 resolution. But, the model's accuracy did not improve. Able to get 0.25 score only. \n\nAny suggestion, how do improve the models performance?",
      "votes": null
    },
    {
      "id": "438320",
      "postDate": "12/13/2018 13:23:33",
      "content": "<p>You can use some of the pointers given in this discussion</p>\n\n<p><a href=\"https://www.kaggle.com/c/human-protein-atlas-image-classification/discussion/73938\">https://www.kaggle.com/c/human-protein-atlas-image-classification/discussion/73938</a></p>",
      "rawMarkdown": "You can use some of the pointers given in this discussion\n\nhttps://www.kaggle.com/c/human-protein-atlas-image-classification/discussion/73938",
      "votes": null
    },
    {
      "id": "439723",
      "postDate": "12/16/2018 07:47:51",
      "content": "<p>Thank you</p>",
      "rawMarkdown": "Thank you",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 438320,
      "author_name": "sandy1112",
      "author_url": "",
      "post_date": "12/13/2018 13:23:33",
      "content": "<p>You can use some of the pointers given in this discussion</p>\n\n<p><a href=\"https://www.kaggle.com/c/human-protein-atlas-image-classification/discussion/73938\">https://www.kaggle.com/c/human-protein-atlas-image-classification/discussion/73938</a></p>",
      "votes": null,
      "replies": [
        {
          "id": 439723,
          "author_name": "anburiyan",
          "author_url": "",
          "post_date": "12/16/2018 07:47:51",
          "content": "<p>Thank you</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "438269": "I tried desnet 169 with the given macro f1 score login. Used 4channels as well 3channels with 512x512 resolution. But, the model's accuracy did not improve. Able to get 0.25 score only. \n\nAny suggestion, how do improve the models performance?",
    "438320": "You can use some of the pointers given in this discussion\n\nhttps://www.kaggle.com/c/human-protein-atlas-image-classification/discussion/73938",
    "439723": "Thank you"
  },
  "source": "meta"
}