{
  "id": 202665,
  "title": "How long does LightGBM takes to train one epoch?",
  "url": "/competitions/riiid-test-answer-prediction/discussion/202665",
  "author_name": "jwc",
  "post_date": "2020-12-11T09:31:56.939000",
  "votes": 2,
  "comment_count": 0,
  "views": 0,
  "content": "<p>I'm using <a href=\"https://www.kaggle.com/its7171/cv-strategy\" target=\"_blank\">this dataset</a> and trying to use <code>cv5_train</code> for training lightgbm model.</p>\n<p>The size of the train dataset is about 88730332,  and it takes about <strong>1 minute in training 1 epoch…!!</strong></p>\n<p>I know that my workstation is not that brand-new but want to know how long is it taken for training 1 epoch of you guys!</p>\n<p>p.s I set <code>n_jobs=-1</code> when training but cpu usage is not always full..  what's wrong with it? Most of the time, only one cpu run 100% and when it is about to go to the next epoch, all cpu usages are getting full just for a second and get back to the one cpu full…</p>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F999983%2F3e3decede280a2464d58a0cc970864c8%2FScreen%20Shot%202020-12-11%20at%2018.28.43.png?generation=1607678959020218&amp;alt=media\" alt=\"\"></p>",
  "messages": [
    {
      "id": 1109054,
      "postDate": "2020-12-11T09:31:56.940Z",
      "content": "<p>I'm using <a href=\"https://www.kaggle.com/its7171/cv-strategy\" target=\"_blank\">this dataset</a> and trying to use <code>cv5_train</code> for training lightgbm model.</p>\n<p>The size of the train dataset is about 88730332,  and it takes about <strong>1 minute in training 1 epoch…!!</strong></p>\n<p>I know that my workstation is not that brand-new but want to know how long is it taken for training 1 epoch of you guys!</p>\n<p>p.s I set <code>n_jobs=-1</code> when training but cpu usage is not always full..  what's wrong with it? Most of the time, only one cpu run 100% and when it is about to go to the next epoch, all cpu usages are getting full just for a second and get back to the one cpu full…</p>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F999983%2F3e3decede280a2464d58a0cc970864c8%2FScreen%20Shot%202020-12-11%20at%2018.28.43.png?generation=1607678959020218&amp;alt=media\" alt=\"\"></p>",
      "rawMarkdown": "I'm using [this dataset](https://www.kaggle.com/its7171/cv-strategy) and trying to use `cv5_train` for training lightgbm model.\n\nThe size of the train dataset is about 88730332,  and it takes about **1 minute in training 1 epoch...!!**\n\nI know that my workstation is not that brand-new but want to know how long is it taken for training 1 epoch of you guys!\n\np.s I set `n_jobs=-1` when training but cpu usage is not always full..  what's wrong with it? Most of the time, only one cpu run 100% and when it is about to go to the next epoch, all cpu usages are getting full just for a second and get back to the one cpu full...\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F999983%2F3e3decede280a2464d58a0cc970864c8%2FScreen%20Shot%202020-12-11%20at%2018.28.43.png?generation=1607678959020218&alt=media)",
      "votes": 2
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1109054": "I'm using [this dataset](https://www.kaggle.com/its7171/cv-strategy) and trying to use `cv5_train` for training lightgbm model.\n\nThe size of the train dataset is about 88730332,  and it takes about **1 minute in training 1 epoch...!!**\n\nI know that my workstation is not that brand-new but want to know how long is it taken for training 1 epoch of you guys!\n\np.s I set `n_jobs=-1` when training but cpu usage is not always full..  what's wrong with it? Most of the time, only one cpu run 100% and when it is about to go to the next epoch, all cpu usages are getting full just for a second and get back to the one cpu full...\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F999983%2F3e3decede280a2464d58a0cc970864c8%2FScreen%20Shot%202020-12-11%20at%2018.28.43.png?generation=1607678959020218&alt=media)"
  }
}