{
  "id": 346264,
  "title": "Is it allowed to make LM with sentences outside the training and validation dataset?",
  "url": "/competitions/dlsprint/discussion/346264",
  "author_name": "",
  "post_date": "2022-08-18T17:37:44.178387300Z",
  "votes": 4,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Is it allowed to make LM using text corpus(obviously open sourced) outside the training and validation dataset? It is necessary for the model to predict the spellings of words not present in the train and validation set correctly as there are several letters of similar pronunciation in Bengali language.</p>",
  "messages": [
    {
      "id": "1905032",
      "postDate": "08/18/2022 17:37:44",
      "content": "<p>Is it allowed to make LM using text corpus(obviously open sourced) outside the training and validation dataset? It is necessary for the model to predict the spellings of words not present in the train and validation set correctly as there are several letters of similar pronunciation in Bengali language.</p>",
      "rawMarkdown": "Is it allowed to make LM using text corpus(obviously open sourced) outside the training and validation dataset? It is necessary for the model to predict the spellings of words not present in the train and validation set correctly as there are several letters of similar pronunciation in Bengali language.",
      "votes": null
    },
    {
      "id": "1905672",
      "postDate": "08/19/2022 08:11:43",
      "content": "<p><a href=\"https://www.kaggle.com/sirajissalakeen\" target=\"_blank\">@sirajissalakeen</a> you mean using lm or spell corrector as a post processor? i think it's allowed to do such post processing , <a href=\"https://www.kaggle.com/sushmit0109\" target=\"_blank\">@sushmit0109</a> <a href=\"https://www.kaggle.com/reasat\" target=\"_blank\">@reasat</a> bhai???</p>",
      "rawMarkdown": "sirajissalakeen you mean using lm or spell corrector as a post processor? i think it's allowed to do such post processing , @sushmit0109 @reasat bhai???",
      "votes": null
    },
    {
      "id": "1905794",
      "postDate": "08/19/2022 09:58:06",
      "content": "<p>Yes, I meant this.</p>",
      "rawMarkdown": "Yes, I meant this.",
      "votes": null
    },
    {
      "id": "1907448",
      "postDate": "08/20/2022 19:43:26",
      "content": "<p><a href=\"https://www.kaggle.com/sushmit0109\" target=\"_blank\">@sushmit0109</a> <a href=\"https://www.kaggle.com/reasat\" target=\"_blank\">@reasat</a> bhai…will you confirm?</p>",
      "rawMarkdown": "sushmit0109 @reasat bhai...will you confirm?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1905672,
      "author_name": "mobassir",
      "author_url": "",
      "post_date": "08/19/2022 08:11:43",
      "content": "<p><a href=\"https://www.kaggle.com/sirajissalakeen\" target=\"_blank\">@sirajissalakeen</a> you mean using lm or spell corrector as a post processor? i think it's allowed to do such post processing , <a href=\"https://www.kaggle.com/sushmit0109\" target=\"_blank\">@sushmit0109</a> <a href=\"https://www.kaggle.com/reasat\" target=\"_blank\">@reasat</a> bhai???</p>",
      "votes": null,
      "replies": [
        {
          "id": 1905794,
          "author_name": "sirajissalakeen",
          "author_url": "",
          "post_date": "08/19/2022 09:58:06",
          "content": "<p>Yes, I meant this.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1907448,
      "author_name": "sawradipsaha",
      "author_url": "",
      "post_date": "08/20/2022 19:43:26",
      "content": "<p><a href=\"https://www.kaggle.com/sushmit0109\" target=\"_blank\">@sushmit0109</a> <a href=\"https://www.kaggle.com/reasat\" target=\"_blank\">@reasat</a> bhai…will you confirm?</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1905032": "Is it allowed to make LM using text corpus(obviously open sourced) outside the training and validation dataset? It is necessary for the model to predict the spellings of words not present in the train and validation set correctly as there are several letters of similar pronunciation in Bengali language.",
    "1905672": "sirajissalakeen you mean using lm or spell corrector as a post processor? i think it's allowed to do such post processing , @sushmit0109 @reasat bhai???",
    "1905794": "Yes, I meant this.",
    "1907448": "sushmit0109 @reasat bhai...will you confirm?"
  },
  "source": "meta"
}