{
  "id": 324124,
  "title": "Comment for begginers",
  "url": "/competitions/birdclef-2022/discussion/324124",
  "author_name": "",
  "post_date": "2022-05-10T08:14:32.618637500Z",
  "votes": 63,
  "comment_count": 15,
  "views": 0,
  "content": "<p>This competition is a sound analysis. <br>\nSince sound competitions are not common on Kaggle, I leave comments for begginers to improve your LB.</p>\n<ul>\n<li><p>Weak label<br>\nThe most important thing is weak label. The best way to learn about weak label is to read <a href=\"https://www.kaggle.com/competitions/birdsong-recognition/discussion/183204\" target=\"_blank\">this discussion</a>.</p></li>\n<li><p>Model architecture<br>\nI recommend <a href=\"https://www.kaggle.com/competitions/rfcx-species-audio-detection/discussion/211007#1151299\" target=\"_blank\">PANNs</a> because this is very powerful.</p></li>\n<li><p>Augmentation<br>\nAugmentation are also important. I wrote about Augmentation <a href=\"https://www.kaggle.com/competitions/birdclef-2022/discussion/307880\" target=\"_blank\">here</a>.</p></li>\n<li><p><a href=\"https://www.kaggle.com/competitions/birdclef-2022/discussion/318999\" target=\"_blank\">Threshold tuning</a></p></li>\n</ul>\n<p>If you do above, you can get <strong>public LB 0.73.</strong></p>",
  "messages": [
    {
      "id": "1783245",
      "postDate": "05/10/2022 08:14:32",
      "content": "<p>This competition is a sound analysis. <br>\nSince sound competitions are not common on Kaggle, I leave comments for begginers to improve your LB.</p>\n<ul>\n<li><p>Weak label<br>\nThe most important thing is weak label. The best way to learn about weak label is to read <a href=\"https://www.kaggle.com/competitions/birdsong-recognition/discussion/183204\" target=\"_blank\">this discussion</a>.</p></li>\n<li><p>Model architecture<br>\nI recommend <a href=\"https://www.kaggle.com/competitions/rfcx-species-audio-detection/discussion/211007#1151299\" target=\"_blank\">PANNs</a> because this is very powerful.</p></li>\n<li><p>Augmentation<br>\nAugmentation are also important. I wrote about Augmentation <a href=\"https://www.kaggle.com/competitions/birdclef-2022/discussion/307880\" target=\"_blank\">here</a>.</p></li>\n<li><p><a href=\"https://www.kaggle.com/competitions/birdclef-2022/discussion/318999\" target=\"_blank\">Threshold tuning</a></p></li>\n</ul>\n<p>If you do above, you can get <strong>public LB 0.73.</strong></p>",
      "rawMarkdown": "This competition is a sound analysis. \nSince sound competitions are not common on Kaggle, I leave comments for begginers to improve your LB.\n\n+ Weak label\nThe most important thing is weak label. The best way to learn about weak label is to read [this discussion](https://www.kaggle.com/competitions/birdsong-recognition/discussion/183204).\n\n+ Model architecture\nI recommend [PANNs](https://www.kaggle.com/competitions/rfcx-species-audio-detection/discussion/211007#1151299) because this is very powerful.\n\n+ Augmentation\nAugmentation are also important. I wrote about Augmentation [here](https://www.kaggle.com/competitions/birdclef-2022/discussion/307880).\n\n+ [Threshold tuning](https://www.kaggle.com/competitions/birdclef-2022/discussion/318999)\n\nIf you do above, you can get **public LB 0.73.**",
      "votes": null
    },
    {
      "id": "1783416",
      "postDate": "05/10/2022 11:42:43",
      "content": "<p>This is helpful. Thanks for sharing!</p>",
      "rawMarkdown": "This is helpful. Thanks for sharing!",
      "votes": null
    },
    {
      "id": "1783569",
      "postDate": "05/10/2022 13:36:20",
      "content": "<p>Thanks for sharing!! You current score is also PANN based?</p>",
      "rawMarkdown": "Thanks for sharing!! You current score is also PANN based?",
      "votes": null
    },
    {
      "id": "1783876",
      "postDate": "05/10/2022 18:50:19",
      "content": "<p>This is a great post for beginners. I would just add that it is also important to try different types of data augmentation and see what works best for your dataset. For example, you could try adding noise, changing the pitch or tempo, etc.</p>",
      "rawMarkdown": "This is a great post for beginners. I would just add that it is also important to try different types of data augmentation and see what works best for your dataset. For example, you could try adding noise, changing the pitch or tempo, etc.",
      "votes": null
    },
    {
      "id": "1784133",
      "postDate": "05/11/2022 00:35:08",
      "content": "<p>This is my first audio competition. It is very helpful post, thanks for sharing. I like this format: idea - link to the context, so I can make my own exploration. Any additional information is very appreciated!</p>",
      "rawMarkdown": "This is my first audio competition. It is very helpful post, thanks for sharing. I like this format: idea - link to the context, so I can make my own exploration. Any additional information is very appreciated!",
      "votes": null
    },
    {
      "id": "1784210",
      "postDate": "05/11/2022 03:00:15",
      "content": "<p>It is a secret.<br>\nBut if you use only PANNs, you will get a silver medal.👍</p>",
      "rawMarkdown": "It is a secret.\nBut if you use only PANNs, you will get a silver medal.👍",
      "votes": null
    },
    {
      "id": "1784219",
      "postDate": "05/11/2022 03:10:26",
      "content": "<p>Thanks for sharing! </p>",
      "rawMarkdown": "Thanks for sharing!",
      "votes": null
    },
    {
      "id": "1784862",
      "postDate": "05/11/2022 14:42:51",
      "content": "<p>Hi my notebook was running fine but the public score is loading and the notebook timeout error comes what Im doing wrong can u help me on this</p>",
      "rawMarkdown": "Hi my notebook was running fine but the public score is loading and the notebook timeout error comes what Im doing wrong can u help me on this",
      "votes": null
    },
    {
      "id": "1784885",
      "postDate": "05/11/2022 15:02:38",
      "content": "<p>Trying to figure out that secret right now as we are only using PANNs right now. 😉</p>",
      "rawMarkdown": "Trying to figure out that secret right now as we are only using PANNs right now. 😉",
      "votes": null
    },
    {
      "id": "1786923",
      "postDate": "05/13/2022 12:54:00",
      "content": "<p>Thanks for sharing !! </p>",
      "rawMarkdown": "Thanks for sharing !!",
      "votes": null
    },
    {
      "id": "1787210",
      "postDate": "05/13/2022 17:00:20",
      "content": "<p>Thanks for sharing :)</p>",
      "rawMarkdown": "Thanks for sharing :)",
      "votes": null
    },
    {
      "id": "1789938",
      "postDate": "05/14/2022 11:11:58",
      "content": "<p>Thanks for sharing :)</p>",
      "rawMarkdown": "Thanks for sharing :)",
      "votes": null
    },
    {
      "id": "1792030",
      "postDate": "05/16/2022 15:17:49",
      "content": "<p>Thanks for sharing :) </p>",
      "rawMarkdown": "Thanks for sharing :)",
      "votes": null
    },
    {
      "id": "1799741",
      "postDate": "05/24/2022 08:59:14",
      "content": "<p>Thanks for sharing :)</p>",
      "rawMarkdown": "Thanks for sharing :)",
      "votes": null
    },
    {
      "id": "1801559",
      "postDate": "05/25/2022 21:44:56",
      "content": "<p>I’ｍ looking forward to your solution.</p>",
      "rawMarkdown": "I’ｍ looking forward to your solution.",
      "votes": null
    },
    {
      "id": "1869116",
      "postDate": "07/24/2022 13:54:06",
      "content": "<p>thanks for sharing</p>",
      "rawMarkdown": "thanks for sharing",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1783416,
      "author_name": "sfnzjc",
      "author_url": "",
      "post_date": "05/10/2022 11:42:43",
      "content": "<p>This is helpful. Thanks for sharing!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1783569,
      "author_name": "bibek777",
      "author_url": "",
      "post_date": "05/10/2022 13:36:20",
      "content": "<p>Thanks for sharing!! You current score is also PANN based?</p>",
      "votes": null,
      "replies": [
        {
          "id": 1784210,
          "author_name": "shinmurashinmura",
          "author_url": "",
          "post_date": "05/11/2022 03:00:15",
          "content": "<p>It is a secret.<br>\nBut if you use only PANNs, you will get a silver medal.👍</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1784885,
          "author_name": "vexxingbanana",
          "author_url": "",
          "post_date": "05/11/2022 15:02:38",
          "content": "<p>Trying to figure out that secret right now as we are only using PANNs right now. 😉</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1783876,
      "author_name": "",
      "author_url": "",
      "post_date": "05/10/2022 18:50:19",
      "content": "<p>This is a great post for beginners. I would just add that it is also important to try different types of data augmentation and see what works best for your dataset. For example, you could try adding noise, changing the pitch or tempo, etc.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1784133,
      "author_name": "azamat25",
      "author_url": "",
      "post_date": "05/11/2022 00:35:08",
      "content": "<p>This is my first audio competition. It is very helpful post, thanks for sharing. I like this format: idea - link to the context, so I can make my own exploration. Any additional information is very appreciated!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1784219,
      "author_name": "yangranran",
      "author_url": "",
      "post_date": "05/11/2022 03:10:26",
      "content": "<p>Thanks for sharing! </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1784862,
      "author_name": "kannanlee",
      "author_url": "",
      "post_date": "05/11/2022 14:42:51",
      "content": "<p>Hi my notebook was running fine but the public score is loading and the notebook timeout error comes what Im doing wrong can u help me on this</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1786923,
      "author_name": "abderrahimnamouh",
      "author_url": "",
      "post_date": "05/13/2022 12:54:00",
      "content": "<p>Thanks for sharing !! </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1787210,
      "author_name": "amitkumar51",
      "author_url": "",
      "post_date": "05/13/2022 17:00:20",
      "content": "<p>Thanks for sharing :)</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1789938,
      "author_name": "saidoulberj",
      "author_url": "",
      "post_date": "05/14/2022 11:11:58",
      "content": "<p>Thanks for sharing :)</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1792030,
      "author_name": "l1uhaif3ng",
      "author_url": "",
      "post_date": "05/16/2022 15:17:49",
      "content": "<p>Thanks for sharing :) </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1799741,
      "author_name": "emreclsr",
      "author_url": "",
      "post_date": "05/24/2022 08:59:14",
      "content": "<p>Thanks for sharing :)</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1801559,
      "author_name": "shigemitsutomizawa",
      "author_url": "",
      "post_date": "05/25/2022 21:44:56",
      "content": "<p>I’ｍ looking forward to your solution.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1869116,
      "author_name": "alaaelsaka",
      "author_url": "",
      "post_date": "07/24/2022 13:54:06",
      "content": "<p>thanks for sharing</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1783245": "This competition is a sound analysis. \nSince sound competitions are not common on Kaggle, I leave comments for begginers to improve your LB.\n\n+ Weak label\nThe most important thing is weak label. The best way to learn about weak label is to read [this discussion](https://www.kaggle.com/competitions/birdsong-recognition/discussion/183204).\n\n+ Model architecture\nI recommend [PANNs](https://www.kaggle.com/competitions/rfcx-species-audio-detection/discussion/211007#1151299) because this is very powerful.\n\n+ Augmentation\nAugmentation are also important. I wrote about Augmentation [here](https://www.kaggle.com/competitions/birdclef-2022/discussion/307880).\n\n+ [Threshold tuning](https://www.kaggle.com/competitions/birdclef-2022/discussion/318999)\n\nIf you do above, you can get **public LB 0.73.**",
    "1783416": "This is helpful. Thanks for sharing!",
    "1783569": "Thanks for sharing!! You current score is also PANN based?",
    "1783876": "This is a great post for beginners. I would just add that it is also important to try different types of data augmentation and see what works best for your dataset. For example, you could try adding noise, changing the pitch or tempo, etc.",
    "1784133": "This is my first audio competition. It is very helpful post, thanks for sharing. I like this format: idea - link to the context, so I can make my own exploration. Any additional information is very appreciated!",
    "1784210": "It is a secret.\nBut if you use only PANNs, you will get a silver medal.👍",
    "1784219": "Thanks for sharing!",
    "1784862": "Hi my notebook was running fine but the public score is loading and the notebook timeout error comes what Im doing wrong can u help me on this",
    "1784885": "Trying to figure out that secret right now as we are only using PANNs right now. 😉",
    "1786923": "Thanks for sharing !!",
    "1787210": "Thanks for sharing :)",
    "1789938": "Thanks for sharing :)",
    "1792030": "Thanks for sharing :)",
    "1799741": "Thanks for sharing :)",
    "1801559": "I’ｍ looking forward to your solution.",
    "1869116": "thanks for sharing"
  },
  "source": "meta"
}