{
  "id": 321126,
  "title": "Best single model and inference time.",
  "url": "/competitions/birdclef-2022/discussion/321126",
  "author_name": "",
  "post_date": "2022-04-25T07:46:57.726303200Z",
  "votes": 3,
  "comment_count": 4,
  "views": 0,
  "content": "<p>As a usual thread: what's your Best Single Model and Inference Time?</p>\n<p>Mine:<br>\nefficientnet_b0<br>\n1 fold<br>\ncv: 0.746<br>\nlb : 0.71<br>\nInference time: 1.5 h</p>",
  "messages": [
    {
      "id": "1767258",
      "postDate": "04/25/2022 07:46:57",
      "content": "<p>As a usual thread: what's your Best Single Model and Inference Time?</p>\n<p>Mine:<br>\nefficientnet_b0<br>\n1 fold<br>\ncv: 0.746<br>\nlb : 0.71<br>\nInference time: 1.5 h</p>",
      "rawMarkdown": "As a usual thread: what's your Best Single Model and Inference Time?\n\nMine:\nefficientnet_b0\n1 fold\ncv: 0.746\nlb : 0.71\nInference time: 1.5 h",
      "votes": null
    },
    {
      "id": "1768375",
      "postDate": "04/26/2022 08:13:41",
      "content": "<p>I achieved 0.72 with a single model, but had to lower the threshold </p>",
      "rawMarkdown": "I achieved 0.72 with a single model, but had to lower the threshold",
      "votes": null
    },
    {
      "id": "1768413",
      "postDate": "04/26/2022 09:13:08",
      "content": "<p>It seems that all people have different validate strategies…..</p>",
      "rawMarkdown": "It seems that all people have different validate strategies.....",
      "votes": null
    },
    {
      "id": "1780195",
      "postDate": "05/07/2022 07:30:59",
      "content": "<p>Mine is 0.73. <a href=\"https://www.kaggle.com/leonshangguan\" target=\"_blank\">@leonshangguan</a> how did you evaluate CV ?</p>",
      "rawMarkdown": "Mine is 0.73. @leonshangguan how did you evaluate CV ?",
      "votes": null
    },
    {
      "id": "1780709",
      "postDate": "05/07/2022 18:06:47",
      "content": "<p>I am using f1_score as stated <a href=\"https://www.kaggle.com/competitions/birdclef-2022/discussion/321871\" target=\"_blank\">metrics.f1_score(np.array(self.y_true), np.array(self.y_pred) &gt; 0.3</a>, this is very close to my lb (+-0.03)</p>",
      "rawMarkdown": "I am using f1_score as stated [metrics.f1_score(np.array(self.y_true), np.array(self.y_pred) > 0.3](https://www.kaggle.com/competitions/birdclef-2022/discussion/321871), this is very close to my lb (+-0.03)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1768375,
      "author_name": "hichambellafkir",
      "author_url": "",
      "post_date": "04/26/2022 08:13:41",
      "content": "<p>I achieved 0.72 with a single model, but had to lower the threshold </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1768413,
      "author_name": "zekunn",
      "author_url": "",
      "post_date": "04/26/2022 09:13:08",
      "content": "<p>It seems that all people have different validate strategies…..</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1780195,
      "author_name": "ulrich07",
      "author_url": "",
      "post_date": "05/07/2022 07:30:59",
      "content": "<p>Mine is 0.73. <a href=\"https://www.kaggle.com/leonshangguan\" target=\"_blank\">@leonshangguan</a> how did you evaluate CV ?</p>",
      "votes": null,
      "replies": [
        {
          "id": 1780709,
          "author_name": "leonshangguan",
          "author_url": "",
          "post_date": "05/07/2022 18:06:47",
          "content": "<p>I am using f1_score as stated <a href=\"https://www.kaggle.com/competitions/birdclef-2022/discussion/321871\" target=\"_blank\">metrics.f1_score(np.array(self.y_true), np.array(self.y_pred) &gt; 0.3</a>, this is very close to my lb (+-0.03)</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1767258": "As a usual thread: what's your Best Single Model and Inference Time?\n\nMine:\nefficientnet_b0\n1 fold\ncv: 0.746\nlb : 0.71\nInference time: 1.5 h",
    "1768375": "I achieved 0.72 with a single model, but had to lower the threshold",
    "1768413": "It seems that all people have different validate strategies.....",
    "1780195": "Mine is 0.73. @leonshangguan how did you evaluate CV ?",
    "1780709": "I am using f1_score as stated [metrics.f1_score(np.array(self.y_true), np.array(self.y_pred) > 0.3](https://www.kaggle.com/competitions/birdclef-2022/discussion/321871), this is very close to my lb (+-0.03)"
  },
  "source": "meta"
}