{
  "id": 392657,
  "title": "Train the model using sound or text",
  "url": "/competitions/ml-olympiad-dialectrecognition/discussion/392657",
  "author_name": "",
  "post_date": "2023-03-06T10:57:17.918966800Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Dear Respected All, <br>\nI hope you are doing great</p>\n<p>My question is how we should train the model using \"audio file\" or \"text\"?<br>\nI mean when we use the data we train using \"GroundTruthText\" and \"SpeakerDialect\" columns or using \"FileName\" and \"SpeakerDialect\" columns.</p>",
  "messages": [
    {
      "id": "2170863",
      "postDate": "03/06/2023 10:57:17",
      "content": "<p>Dear Respected All, <br>\nI hope you are doing great</p>\n<p>My question is how we should train the model using \"audio file\" or \"text\"?<br>\nI mean when we use the data we train using \"GroundTruthText\" and \"SpeakerDialect\" columns or using \"FileName\" and \"SpeakerDialect\" columns.</p>",
      "rawMarkdown": "Dear Respected All, \nI hope you are doing great\n\nMy question is how we should train the model using \"audio file\" or \"text\"?\nI mean when we use the data we train using \"GroundTruthText\" and \"SpeakerDialect\" columns or using \"FileName\" and \"SpeakerDialect\" columns.",
      "votes": null
    },
    {
      "id": "2170957",
      "postDate": "03/06/2023 12:22:56",
      "content": "<p>Hello Mohammed,</p>\n<p>I believe the objective of this competition is dialect classification, you can use either the text or audio to accomplish that, however for the audio you have to read the wav file and prepare the input data (check the Segment Start, End, Length, and source Filename).<br>\nFor the text classification, check <a href=\"https://www.kaggle.com/code/asalhi/starter-training-and-infer-using-arabert\" target=\"_blank\">ALI SALHI's</a> notebook it provides a good start.</p>",
      "rawMarkdown": "Hello Mohammed,\n\nI believe the objective of this competition is dialect classification, you can use either the text or audio to accomplish that, however for the audio you have to read the wav file and prepare the input data (check the Segment Start, End, Length, and source Filename).\nFor the text classification, check [ALI SALHI's](https://www.kaggle.com/code/asalhi/starter-training-and-infer-using-arabert) notebook it provides a good start.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2170957,
      "author_name": "salimabdelaziz",
      "author_url": "",
      "post_date": "03/06/2023 12:22:56",
      "content": "<p>Hello Mohammed,</p>\n<p>I believe the objective of this competition is dialect classification, you can use either the text or audio to accomplish that, however for the audio you have to read the wav file and prepare the input data (check the Segment Start, End, Length, and source Filename).<br>\nFor the text classification, check <a href=\"https://www.kaggle.com/code/asalhi/starter-training-and-infer-using-arabert\" target=\"_blank\">ALI SALHI's</a> notebook it provides a good start.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2170863": "Dear Respected All, \nI hope you are doing great\n\nMy question is how we should train the model using \"audio file\" or \"text\"?\nI mean when we use the data we train using \"GroundTruthText\" and \"SpeakerDialect\" columns or using \"FileName\" and \"SpeakerDialect\" columns.",
    "2170957": "Hello Mohammed,\n\nI believe the objective of this competition is dialect classification, you can use either the text or audio to accomplish that, however for the audio you have to read the wav file and prepare the input data (check the Segment Start, End, Length, and source Filename).\nFor the text classification, check [ALI SALHI's](https://www.kaggle.com/code/asalhi/starter-training-and-infer-using-arabert) notebook it provides a good start."
  },
  "source": "meta"
}