{
  "id": 291694,
  "title": "Customize your Faster RCNN",
  "url": "/competitions/tensorflow-great-barrier-reef/discussion/291694",
  "author_name": "",
  "post_date": "2021-11-30T15:36:50.787612Z",
  "votes": 20,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Here is nice article how to technicaly customize FasterRCNN with or without FPN with different heads , losses etc,</p>\n<p><a href=\"https://akashprakas.github.io/My-blog/jupyter/2020/12/19/Hacking_fasterRcnn.html\" target=\"_blank\">https://akashprakas.github.io/My-blog/jupyter/2020/12/19/Hacking_fasterRcnn.html</a></p>",
  "messages": [
    {
      "id": "1600649",
      "postDate": "11/30/2021 15:36:50",
      "content": "<p>Here is nice article how to technicaly customize FasterRCNN with or without FPN with different heads , losses etc,</p>\n<p><a href=\"https://akashprakas.github.io/My-blog/jupyter/2020/12/19/Hacking_fasterRcnn.html\" target=\"_blank\">https://akashprakas.github.io/My-blog/jupyter/2020/12/19/Hacking_fasterRcnn.html</a></p>",
      "rawMarkdown": "Here is nice article how to technicaly customize FasterRCNN with or without FPN with different heads , losses etc,\n\nhttps://akashprakas.github.io/My-blog/jupyter/2020/12/19/Hacking_fasterRcnn.html",
      "votes": null
    },
    {
      "id": "1600999",
      "postDate": "11/30/2021 22:55:12",
      "content": "<p>Great post! Thank you Łukasz! It is great for many reasons - easy to understand architecture, commonly used, Pytorch, exhaustive explanations and many tricks and tips. </p>",
      "rawMarkdown": "Great post! Thank you Łukasz! It is great for many reasons - easy to understand architecture, commonly used, Pytorch, exhaustive explanations and many tricks and tips.",
      "votes": null
    },
    {
      "id": "1601235",
      "postDate": "12/01/2021 05:38:57",
      "content": "<p>Your Welcome Remek, found it usefull so i shared. As far as i am testing the aspect_ratios from anchor boxes got also impact on score :)</p>",
      "rawMarkdown": "Your Welcome Remek, found it usefull so i shared. As far as i am testing the aspect_ratios from anchor boxes got also impact on score :)",
      "votes": null
    },
    {
      "id": "1602062",
      "postDate": "12/01/2021 18:21:37",
      "content": "<p>Wow! this is great. Thank you for sharing the blog post.</p>",
      "rawMarkdown": "Wow! this is great. Thank you for sharing the blog post.",
      "votes": null
    },
    {
      "id": "1608207",
      "postDate": "12/06/2021 09:45:00",
      "content": "<p>Thanks for sharing <a href=\"https://www.kaggle.com/remekkinas\" target=\"_blank\">@remekkinas</a> , this is great!</p>",
      "rawMarkdown": "Thanks for sharing @remekkinas , this is great!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1600999,
      "author_name": "remekkinas",
      "author_url": "",
      "post_date": "11/30/2021 22:55:12",
      "content": "<p>Great post! Thank you Łukasz! It is great for many reasons - easy to understand architecture, commonly used, Pytorch, exhaustive explanations and many tricks and tips. </p>",
      "votes": null,
      "replies": [
        {
          "id": 1601235,
          "author_name": "lukaszborecki",
          "author_url": "",
          "post_date": "12/01/2021 05:38:57",
          "content": "<p>Your Welcome Remek, found it usefull so i shared. As far as i am testing the aspect_ratios from anchor boxes got also impact on score :)</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1602062,
      "author_name": "soumya9977",
      "author_url": "",
      "post_date": "12/01/2021 18:21:37",
      "content": "<p>Wow! this is great. Thank you for sharing the blog post.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1608207,
      "author_name": "saurabhbagchi",
      "author_url": "",
      "post_date": "12/06/2021 09:45:00",
      "content": "<p>Thanks for sharing <a href=\"https://www.kaggle.com/remekkinas\" target=\"_blank\">@remekkinas</a> , this is great!</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1600649": "Here is nice article how to technicaly customize FasterRCNN with or without FPN with different heads , losses etc,\n\nhttps://akashprakas.github.io/My-blog/jupyter/2020/12/19/Hacking_fasterRcnn.html",
    "1600999": "Great post! Thank you Łukasz! It is great for many reasons - easy to understand architecture, commonly used, Pytorch, exhaustive explanations and many tricks and tips.",
    "1601235": "Your Welcome Remek, found it usefull so i shared. As far as i am testing the aspect_ratios from anchor boxes got also impact on score :)",
    "1602062": "Wow! this is great. Thank you for sharing the blog post.",
    "1608207": "Thanks for sharing @remekkinas , this is great!"
  },
  "source": "meta"
}