{
  "id": 451721,
  "title": "119 rank solution ",
  "url": "/competitions/bengaliai-speech/discussion/451721",
  "author_name": "Abheeshth Mishra",
  "post_date": "2023-10-30T08:55:01.789000",
  "votes": 10,
  "comment_count": 0,
  "views": 0,
  "content": "<p>First of all Thanks to kaggle for organising this event . My solution is nothing fancy cause i joined late and couldn't give much time to it(Sadly)</p>\n<p>Model:<br>\nI have used this notebook for my solution - <a href=\"https://www.kaggle.com/code/goelyash/bengali-ai-wave-2-vec-2\" target=\"_blank\">https://www.kaggle.com/code/goelyash/bengali-ai-wave-2-vec-2</a><br>\nChanged some paramters in that notebook . added parameters that is mentioned in this notebook --&gt; <a href=\"https://www.kaggle.com/code/dangnguyen97/lb-0-443-change-the-decoding-parameters-vote-up\" target=\"_blank\">https://www.kaggle.com/code/dangnguyen97/lb-0-443-change-the-decoding-parameters-vote-up</a></p>\n<p>I wish i could play more with parameters  cause i got to know that just by changing the parameters we could have get the good results.</p>\n<p>Things didn't work for me:<br>\nI tried to train the model from scratch but couldn't train for more than 40K steps and the results from 40k checkpoint was around 0.440<br>\nI have also tried to create my own 5gram model but it didn't work for me that well(dont know why )</p>\n<p>congratulations to those who got medals and good ranking in this comp😁🥳</p>",
  "messages": [
    {
      "id": 2504920,
      "postDate": "2023-10-30T08:55:01.790Z",
      "content": "<p>First of all Thanks to kaggle for organising this event . My solution is nothing fancy cause i joined late and couldn't give much time to it(Sadly)</p>\n<p>Model:<br>\nI have used this notebook for my solution - <a href=\"https://www.kaggle.com/code/goelyash/bengali-ai-wave-2-vec-2\" target=\"_blank\">https://www.kaggle.com/code/goelyash/bengali-ai-wave-2-vec-2</a><br>\nChanged some paramters in that notebook . added parameters that is mentioned in this notebook --&gt; <a href=\"https://www.kaggle.com/code/dangnguyen97/lb-0-443-change-the-decoding-parameters-vote-up\" target=\"_blank\">https://www.kaggle.com/code/dangnguyen97/lb-0-443-change-the-decoding-parameters-vote-up</a></p>\n<p>I wish i could play more with parameters  cause i got to know that just by changing the parameters we could have get the good results.</p>\n<p>Things didn't work for me:<br>\nI tried to train the model from scratch but couldn't train for more than 40K steps and the results from 40k checkpoint was around 0.440<br>\nI have also tried to create my own 5gram model but it didn't work for me that well(dont know why )</p>\n<p>congratulations to those who got medals and good ranking in this comp😁🥳</p>",
      "rawMarkdown": "First of all Thanks to kaggle for organising this event . My solution is nothing fancy cause i joined late and couldn't give much time to it(Sadly)\n\nModel:\nI have used this notebook for my solution - https://www.kaggle.com/code/goelyash/bengali-ai-wave-2-vec-2\nChanged some paramters in that notebook . added parameters that is mentioned in this notebook --> https://www.kaggle.com/code/dangnguyen97/lb-0-443-change-the-decoding-parameters-vote-up\n\nI wish i could play more with parameters  cause i got to know that just by changing the parameters we could have get the good results.\n\nThings didn't work for me:\nI tried to train the model from scratch but couldn't train for more than 40K steps and the results from 40k checkpoint was around 0.440\nI have also tried to create my own 5gram model but it didn't work for me that well(dont know why )\n\ncongratulations to those who got medals and good ranking in this comp😁🥳\n",
      "votes": 10
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "2504920": "First of all Thanks to kaggle for organising this event . My solution is nothing fancy cause i joined late and couldn't give much time to it(Sadly)\n\nModel:\nI have used this notebook for my solution - https://www.kaggle.com/code/goelyash/bengali-ai-wave-2-vec-2\nChanged some paramters in that notebook . added parameters that is mentioned in this notebook --> https://www.kaggle.com/code/dangnguyen97/lb-0-443-change-the-decoding-parameters-vote-up\n\nI wish i could play more with parameters  cause i got to know that just by changing the parameters we could have get the good results.\n\nThings didn't work for me:\nI tried to train the model from scratch but couldn't train for more than 40K steps and the results from 40k checkpoint was around 0.440\nI have also tried to create my own 5gram model but it didn't work for me that well(dont know why )\n\ncongratulations to those who got medals and good ranking in this comp😁🥳\n"
  }
}