{
  "id": 486221,
  "title": "Notebook out of memory when scoring",
  "url": "/competitions/home-credit-credit-risk-model-stability/discussion/486221",
  "author_name": "",
  "post_date": "2024-03-24T03:33:00.808716700Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>In view of the public test data only takes up 30% of the whole test data, whether there is such a situation -- a model that works properly on the public test data will run out of memory on the private test data in the final phase.</p>",
  "messages": [
    {
      "id": "2713175",
      "postDate": "03/24/2024 03:33:00",
      "content": "<p>In view of the public test data only takes up 30% of the whole test data, whether there is such a situation -- a model that works properly on the public test data will run out of memory on the private test data in the final phase.</p>",
      "rawMarkdown": "In view of the public test data only takes up 30% of the whole test data, whether there is such a situation -- a model that works properly on the public test data will run out of memory on the private test data in the final phase.",
      "votes": null
    },
    {
      "id": "2713448",
      "postDate": "03/24/2024 07:57:29",
      "content": "<p>Hi,</p>\n<p>Kaggle scores on the whole dataset and only show you the results of the split with 30% of the data, but the other split with the remaining 70% is also scored when you submit, so you don't need to worry about that.</p>",
      "rawMarkdown": "Hi,\n\nKaggle scores on the whole dataset and only show you the results of the split with 30% of the data, but the other split with the remaining 70% is also scored when you submit, so you don't need to worry about that.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2713448,
      "author_name": "enriquezaf",
      "author_url": "",
      "post_date": "03/24/2024 07:57:29",
      "content": "<p>Hi,</p>\n<p>Kaggle scores on the whole dataset and only show you the results of the split with 30% of the data, but the other split with the remaining 70% is also scored when you submit, so you don't need to worry about that.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2713175": "In view of the public test data only takes up 30% of the whole test data, whether there is such a situation -- a model that works properly on the public test data will run out of memory on the private test data in the final phase.",
    "2713448": "Hi,\n\nKaggle scores on the whole dataset and only show you the results of the split with 30% of the data, but the other split with the remaining 70% is also scored when you submit, so you don't need to worry about that."
  },
  "source": "meta"
}