{
  "id": 242453,
  "title": "Grab a coffee but waited too long - 8 hours to having the result",
  "url": "/competitions/coleridgeinitiative-show-us-the-data/discussion/242453",
  "author_name": "",
  "post_date": "2021-05-29T04:14:41.105480700Z",
  "votes": 2,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Hi all,<br>\nIt takes me around 7.5-8.5 hours to completing my submission. I shorten text inputs by taking only sentences that contain words: data, dataset, database, sample, and sentences with its word count is greater than 7 (100 train *.json have 3200 sentences and take 6 minutes being finished in getting dataset name by a bert model), it is still large inputs. How you could reduce your computing time?</p>",
  "messages": [
    {
      "id": "1327204",
      "postDate": "05/29/2021 04:14:41",
      "content": "<p>Hi all,<br>\nIt takes me around 7.5-8.5 hours to completing my submission. I shorten text inputs by taking only sentences that contain words: data, dataset, database, sample, and sentences with its word count is greater than 7 (100 train *.json have 3200 sentences and take 6 minutes being finished in getting dataset name by a bert model), it is still large inputs. How you could reduce your computing time?</p>",
      "rawMarkdown": "Hi all,\n\nIt takes me around 7.5-8.5 hours to completing my submission. I shorten text inputs by taking only sentences that contain words: data, dataset, database, sample, and sentences with its word count is greater than 7 (100 train *.json have 3200 sentences and take 6 minutes being finished in getting dataset name by a bert model), it is still large inputs. How you could reduce your computing time?",
      "votes": null
    },
    {
      "id": "1327528",
      "postDate": "05/29/2021 11:46:39",
      "content": "<p>I have had the same issue</p>",
      "rawMarkdown": "I have had the same issue",
      "votes": null
    },
    {
      "id": "1327755",
      "postDate": "05/29/2021 15:47:19",
      "content": "<p>yes, when I take sentences that contains \"data\" only, my score is 0.33, 6 hours finished, when I add more \"database\", \"dataset\", \"sample\" score is 0.386 but 7-8 hours finished, so score increases increases if the number of sentence inputs increases, at a threshold of filter words number, my notebook run out of time. I am trying to solve speed first maybe use lower embedding dimension like distibert or albert, etc,</p>",
      "rawMarkdown": "yes, when I take sentences that contains \"data\" only, my score is 0.33, 6 hours finished, when I add more \"database\", \"dataset\", \"sample\" score is 0.386 but 7-8 hours finished, so score increases increases if the number of sentence inputs increases, at a threshold of filter words number, my notebook run out of time. I am trying to solve speed first maybe use lower embedding dimension like distibert or albert, etc,",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1327528,
      "author_name": "domdejonge",
      "author_url": "",
      "post_date": "05/29/2021 11:46:39",
      "content": "<p>I have had the same issue</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1327755,
      "author_name": "nguyenvanlinh",
      "author_url": "",
      "post_date": "05/29/2021 15:47:19",
      "content": "<p>yes, when I take sentences that contains \"data\" only, my score is 0.33, 6 hours finished, when I add more \"database\", \"dataset\", \"sample\" score is 0.386 but 7-8 hours finished, so score increases increases if the number of sentence inputs increases, at a threshold of filter words number, my notebook run out of time. I am trying to solve speed first maybe use lower embedding dimension like distibert or albert, etc,</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1327204": "Hi all,\n\nIt takes me around 7.5-8.5 hours to completing my submission. I shorten text inputs by taking only sentences that contain words: data, dataset, database, sample, and sentences with its word count is greater than 7 (100 train *.json have 3200 sentences and take 6 minutes being finished in getting dataset name by a bert model), it is still large inputs. How you could reduce your computing time?",
    "1327528": "I have had the same issue",
    "1327755": "yes, when I take sentences that contains \"data\" only, my score is 0.33, 6 hours finished, when I add more \"database\", \"dataset\", \"sample\" score is 0.386 but 7-8 hours finished, so score increases increases if the number of sentence inputs increases, at a threshold of filter words number, my notebook run out of time. I am trying to solve speed first maybe use lower embedding dimension like distibert or albert, etc,"
  },
  "source": "meta"
}