{
  "id": 337729,
  "title": "Slow notebook",
  "url": "/competitions/amex-default-prediction/discussion/337729",
  "author_name": "",
  "post_date": "2022-07-17T10:44:22.115653600Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Any way to speed up kaggle notebooks ? It takes lot of time for even simple operation like converting from string to datetime. Please ignore if already discussed, newbie here.</p>",
  "messages": [
    {
      "id": "1859009",
      "postDate": "07/17/2022 10:44:22",
      "content": "<p>Any way to speed up kaggle notebooks ? It takes lot of time for even simple operation like converting from string to datetime. Please ignore if already discussed, newbie here.</p>",
      "rawMarkdown": "Any way to speed up kaggle notebooks ? It takes lot of time for even simple operation like converting from string to datetime. Please ignore if already discussed, newbie here.",
      "votes": null
    },
    {
      "id": "1859045",
      "postDate": "07/17/2022 11:34:05",
      "content": "<p>The sheer data size could cause this. You may perhaps consider using the smaller version in parquet/ feather. I think some Kaggle users have truncated the size of the data for everyone's benefit. </p>",
      "rawMarkdown": "The sheer data size could cause this. You may perhaps consider using the smaller version in parquet/ feather. I think some Kaggle users have truncated the size of the data for everyone's benefit.",
      "votes": null
    },
    {
      "id": "1859397",
      "postDate": "07/17/2022 15:38:37",
      "content": "<p>If you perform dataframe operations on the GPU instead of CPU, you can achieve a huge speed up. I show an example <a href=\"https://www.kaggle.com/code/cdeotte/xgboost-starter-0-793\" target=\"_blank\">here</a> using RAPIDS cudf.</p>",
      "rawMarkdown": "If you perform dataframe operations on the GPU instead of CPU, you can achieve a huge speed up. I show an example [here][1] using RAPIDS cudf.\n\n[1]: https://www.kaggle.com/code/cdeotte/xgboost-starter-0-793",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1859045,
      "author_name": "ravi20076",
      "author_url": "",
      "post_date": "07/17/2022 11:34:05",
      "content": "<p>The sheer data size could cause this. You may perhaps consider using the smaller version in parquet/ feather. I think some Kaggle users have truncated the size of the data for everyone's benefit. </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1859397,
      "author_name": "cdeotte",
      "author_url": "",
      "post_date": "07/17/2022 15:38:37",
      "content": "<p>If you perform dataframe operations on the GPU instead of CPU, you can achieve a huge speed up. I show an example <a href=\"https://www.kaggle.com/code/cdeotte/xgboost-starter-0-793\" target=\"_blank\">here</a> using RAPIDS cudf.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1859009": "Any way to speed up kaggle notebooks ? It takes lot of time for even simple operation like converting from string to datetime. Please ignore if already discussed, newbie here.",
    "1859045": "The sheer data size could cause this. You may perhaps consider using the smaller version in parquet/ feather. I think some Kaggle users have truncated the size of the data for everyone's benefit.",
    "1859397": "If you perform dataframe operations on the GPU instead of CPU, you can achieve a huge speed up. I show an example [here][1] using RAPIDS cudf.\n\n[1]: https://www.kaggle.com/code/cdeotte/xgboost-starter-0-793"
  },
  "source": "meta"
}