{
  "id": 71757,
  "title": "How to improve the running speed?",
  "url": "/competitions/quora-insincere-questions-classification/discussion/71757",
  "author_name": "",
  "post_date": "2018-11-16T08:25:05.612418Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>load 4 emb\nuse 2 CuDNNLSTM with units=128,\nI want to try KFold, But the operation lasted for more than 2 hours\nIs there a faster way to load emb?\nHow to make model run faster?\nIn my code,units=128 get Better results compared with  units=64</p>",
  "messages": [
    {
      "id": "422445",
      "postDate": "11/16/2018 08:25:05",
      "content": "<p>load 4 emb\nuse 2 CuDNNLSTM with units=128,\nI want to try KFold, But the operation lasted for more than 2 hours\nIs there a faster way to load emb?\nHow to make model run faster?\nIn my code,units=128 get Better results compared with  units=64</p>",
      "rawMarkdown": "load 4 emb\nuse 2 CuDNNLSTM with units=128,\nI want to try KFold, But the operation lasted for more than 2 hours\nIs there a faster way to load emb?\nHow to make model run faster?\nIn my code,units=128 get Better results compared with  units=64",
      "votes": null
    },
    {
      "id": "422481",
      "postDate": "11/16/2018 09:30:58",
      "content": "<p>In my implementation, reduce maximum length can be help a bit.</p>",
      "rawMarkdown": "In my implementation, reduce maximum length can be help a bit.",
      "votes": null
    },
    {
      "id": "422483",
      "postDate": "11/16/2018 09:33:02",
      "content": "<p>I just use 40</p>",
      "rawMarkdown": "I just use 40",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 422481,
      "author_name": "ratthachat",
      "author_url": "",
      "post_date": "11/16/2018 09:30:58",
      "content": "<p>In my implementation, reduce maximum length can be help a bit.</p>",
      "votes": null,
      "replies": [
        {
          "id": 422483,
          "author_name": "tianshuaifei",
          "author_url": "",
          "post_date": "11/16/2018 09:33:02",
          "content": "<p>I just use 40</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "422445": "load 4 emb\nuse 2 CuDNNLSTM with units=128,\nI want to try KFold, But the operation lasted for more than 2 hours\nIs there a faster way to load emb?\nHow to make model run faster?\nIn my code,units=128 get Better results compared with  units=64",
    "422481": "In my implementation, reduce maximum length can be help a bit.",
    "422483": "I just use 40"
  },
  "source": "meta"
}