{
  "id": 501210,
  "title": "time about scoring",
  "url": "/competitions/home-credit-credit-risk-model-stability/discussion/501210",
  "author_name": "",
  "post_date": "2024-05-08T13:20:49.219235Z",
  "votes": 1,
  "comment_count": 4,
  "views": 0,
  "content": "<p>I'm a freshman, and I'd like to ask how long it usually takes to get a score after submitting it.</p>",
  "messages": [
    {
      "id": "2801079",
      "postDate": "05/08/2024 13:20:49",
      "content": "<p>I'm a freshman, and I'd like to ask how long it usually takes to get a score after submitting it.</p>",
      "rawMarkdown": "I'm a freshman, and I'd like to ask how long it usually takes to get a score after submitting it.",
      "votes": null
    },
    {
      "id": "2801094",
      "postDate": "05/08/2024 13:31:45",
      "content": "<p>It largely depends on the number of models and complexity. You can choose to submit a single model and expect a score in 30-60 minutes for LGBM and about 3-4 hours for catboost (training + inference). Only inference is much shorter and you can expect a single model response in 30 minutes or lesser. An ensemble will obviously take more time <a href=\"https://www.kaggle.com/zexiuw\" target=\"_blank\">@zexiuw</a> </p>",
      "rawMarkdown": "It largely depends on the number of models and complexity. You can choose to submit a single model and expect a score in 30-60 minutes for LGBM and about 3-4 hours for catboost (training + inference). Only inference is much shorter and you can expect a single model response in 30 minutes or lesser. An ensemble will obviously take more time @zexiuw",
      "votes": null
    },
    {
      "id": "2801377",
      "postDate": "05/08/2024 15:39:37",
      "content": "<p>Thanks for your answer！🙂</p>",
      "rawMarkdown": "Thanks for your answer！🙂",
      "votes": null
    },
    {
      "id": "2802903",
      "postDate": "05/09/2024 08:16:27",
      "content": "<p>Respected statified k fold split each round with lightgbm &amp; catbooting model training &amp; inference taking about 2000-3000 seconds.</p>",
      "rawMarkdown": "Respected statified k fold split each round with lightgbm & catbooting model training & inference taking about 2000-3000 seconds.",
      "votes": null
    },
    {
      "id": "2803354",
      "postDate": "05/09/2024 13:02:07",
      "content": "<p>My suggestion is pre-training + inference, because inference can get scores in a very short time, you can start from the XGB and LGBM model to try, the model is relatively simple, the training time relative to Cat can save a lot of time!</p>",
      "rawMarkdown": "My suggestion is pre-training + inference, because inference can get scores in a very short time, you can start from the XGB and LGBM model to try, the model is relatively simple, the training time relative to Cat can save a lot of time!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2801094,
      "author_name": "ravi20076",
      "author_url": "",
      "post_date": "05/08/2024 13:31:45",
      "content": "<p>It largely depends on the number of models and complexity. You can choose to submit a single model and expect a score in 30-60 minutes for LGBM and about 3-4 hours for catboost (training + inference). Only inference is much shorter and you can expect a single model response in 30 minutes or lesser. An ensemble will obviously take more time <a href=\"https://www.kaggle.com/zexiuw\" target=\"_blank\">@zexiuw</a> </p>",
      "votes": null,
      "replies": [
        {
          "id": 2801377,
          "author_name": "zexiuw",
          "author_url": "",
          "post_date": "05/08/2024 15:39:37",
          "content": "<p>Thanks for your answer！🙂</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 2802903,
      "author_name": "kunduruanil",
      "author_url": "",
      "post_date": "05/09/2024 08:16:27",
      "content": "<p>Respected statified k fold split each round with lightgbm &amp; catbooting model training &amp; inference taking about 2000-3000 seconds.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2803354,
      "author_name": "majiaqi111",
      "author_url": "",
      "post_date": "05/09/2024 13:02:07",
      "content": "<p>My suggestion is pre-training + inference, because inference can get scores in a very short time, you can start from the XGB and LGBM model to try, the model is relatively simple, the training time relative to Cat can save a lot of time!</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2801079": "I'm a freshman, and I'd like to ask how long it usually takes to get a score after submitting it.",
    "2801094": "It largely depends on the number of models and complexity. You can choose to submit a single model and expect a score in 30-60 minutes for LGBM and about 3-4 hours for catboost (training + inference). Only inference is much shorter and you can expect a single model response in 30 minutes or lesser. An ensemble will obviously take more time @zexiuw",
    "2801377": "Thanks for your answer！🙂",
    "2802903": "Respected statified k fold split each round with lightgbm & catbooting model training & inference taking about 2000-3000 seconds.",
    "2803354": "My suggestion is pre-training + inference, because inference can get scores in a very short time, you can start from the XGB and LGBM model to try, the model is relatively simple, the training time relative to Cat can save a lot of time!"
  },
  "source": "meta"
}