{
  "id": 461072,
  "title": "184th Place Solution ESM2 (2 models) using RMDB + QUICK_START datasets",
  "url": "/competitions/stanford-ribonanza-rna-folding/discussion/461072",
  "author_name": "Pranshu Bahadur",
  "post_date": "2023-12-12T14:10:41.689000",
  "votes": 2,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Hey everyone,</p>\n<p>Hope y'all had a good time! This was my first competition and I got lucky after the LB shake-up haha (top-68% public -&gt; top-25% private). Super excited for competitions like this in the future.</p>\n<p>Here's a link to my solution - if you are interested in seeing how finetuned ESM2 performs on RMDB + QUICK_START Dataset using 2 models (1 for each experiment type):</p>\n<p><a href=\"https://www.kaggle.com/code/pranshubahadur/esm2-rmdb-rna-dataset\" target=\"_blank\">https://www.kaggle.com/code/pranshubahadur/esm2-rmdb-rna-dataset</a></p>\n<p>This was a regression problem in predicting the reactivity of RNA sequences. Needless to say, I learned a lot about NLP while solving this problem statement. </p>\n<p>It could've been much better if I had explored the bpp files imo…I will learn from my mistakes and improve my approach in the future…</p>\n<p>Be sure to follow  / upvote if you like my work &amp; it brought some value to you!</p>\n<p>Thank you for the opportunity hosts!</p>\n<p>Congratulations to the winners!</p>",
  "messages": [
    {
      "id": 2558986,
      "postDate": "2023-12-12T14:10:41.690Z",
      "content": "<p>Hey everyone,</p>\n<p>Hope y'all had a good time! This was my first competition and I got lucky after the LB shake-up haha (top-68% public -&gt; top-25% private). Super excited for competitions like this in the future.</p>\n<p>Here's a link to my solution - if you are interested in seeing how finetuned ESM2 performs on RMDB + QUICK_START Dataset using 2 models (1 for each experiment type):</p>\n<p><a href=\"https://www.kaggle.com/code/pranshubahadur/esm2-rmdb-rna-dataset\" target=\"_blank\">https://www.kaggle.com/code/pranshubahadur/esm2-rmdb-rna-dataset</a></p>\n<p>This was a regression problem in predicting the reactivity of RNA sequences. Needless to say, I learned a lot about NLP while solving this problem statement. </p>\n<p>It could've been much better if I had explored the bpp files imo…I will learn from my mistakes and improve my approach in the future…</p>\n<p>Be sure to follow  / upvote if you like my work &amp; it brought some value to you!</p>\n<p>Thank you for the opportunity hosts!</p>\n<p>Congratulations to the winners!</p>",
      "rawMarkdown": "Hey everyone,\n\nHope y'all had a good time! This was my first competition and I got lucky after the LB shake-up haha (top-68% public -> top-25% private). Super excited for competitions like this in the future.\n\nHere's a link to my solution - if you are interested in seeing how finetuned ESM2 performs on RMDB + QUICK_START Dataset using 2 models (1 for each experiment type):\n\nhttps://www.kaggle.com/code/pranshubahadur/esm2-rmdb-rna-dataset\n\nThis was a regression problem in predicting the reactivity of RNA sequences. Needless to say, I learned a lot about NLP while solving this problem statement. \n\nIt could've been much better if I had explored the bpp files imo...I will learn from my mistakes and improve my approach in the future...\n\nBe sure to follow  / upvote if you like my work & it brought some value to you!\n\nThank you for the opportunity hosts!\n\nCongratulations to the winners!",
      "votes": 2
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "2558986": "Hey everyone,\n\nHope y'all had a good time! This was my first competition and I got lucky after the LB shake-up haha (top-68% public -> top-25% private). Super excited for competitions like this in the future.\n\nHere's a link to my solution - if you are interested in seeing how finetuned ESM2 performs on RMDB + QUICK_START Dataset using 2 models (1 for each experiment type):\n\nhttps://www.kaggle.com/code/pranshubahadur/esm2-rmdb-rna-dataset\n\nThis was a regression problem in predicting the reactivity of RNA sequences. Needless to say, I learned a lot about NLP while solving this problem statement. \n\nIt could've been much better if I had explored the bpp files imo...I will learn from my mistakes and improve my approach in the future...\n\nBe sure to follow  / upvote if you like my work & it brought some value to you!\n\nThank you for the opportunity hosts!\n\nCongratulations to the winners!"
  }
}