🐦 Twitter Post Details

Viewing enriched Twitter post

@virattt

I finetuned another LLM on financial Q&A. From scratch. Implementation details: • 355M param LLM • 6K training samples • 0.67 training loss • 0.90 validation loss I used a single A100 and it took ~7 minutes. Really cool to see the before and after results. Before: LLM generates random text. After: LLM generates an answer attempt. Shout out to @rasbt for the code.

Media 1

šŸ“Š Media Metadata

{
  "media": [
    {
      "url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/1844132160195768579/media_0.jpg",
      "type": "photo",
      "original_url": "https://pbs.twimg.com/media/GZesbOBWQAAaLK9.jpg",
      "download_date": "2025-08-13T05:53:17.879241",
      "stored_in_supabase": true
    }
  ],
  "conversion_date": "2025-08-13T00:38:59.256356",
  "format_converted": true,
  "original_structure": "had_media_only"
}

šŸ”§ Raw API Response

{
  "user": {
    "created_at": "2010-04-25T19:12:53.000Z",
    "default_profile_image": false,
    "description": "playing with gen ai + finance",
    "fast_followers_count": 0,
    "favourites_count": 7457,
    "followers_count": 17824,
    "friends_count": 78,
    "has_custom_timelines": true,
    "is_translator": false,
    "listed_count": 286,
    "location": "New York, NY",
    "media_count": 549,
    "name": "virat",
    "normal_followers_count": 17824,
    "possibly_sensitive": false,
    "profile_banner_url": "https://pbs.twimg.com/profile_banners/137086701/1678548959",
    "profile_image_url_https": "https://pbs.twimg.com/profile_images/1237964548704841736/iai1MNLE_normal.jpg",
    "screen_name": "virattt",
    "statuses_count": 5785,
    "translator_type": "none",
    "url": "https://t.co/ZTELBmcX06",
    "verified": true,
    "withheld_in_countries": [],
    "id_str": "137086701"
  },
  "id": "1844132160195768579",
  "conversation_id": "1844132160195768579",
  "full_text": "I finetuned another LLM on financial Q&A.\n\nFrom scratch.\n\nImplementation details:\n• 355M param LLM\n• 6K training samples\n• 0.67 training loss\n• 0.90 validation loss\n\nI used a single A100 and it took ~7 minutes.\n\nReally cool to see the before and after results.\n\nBefore: LLM generates random text.\n\nAfter: LLM generates an answer attempt.\n\nShout out to @rasbt for the code.",
  "reply_count": 18,
  "retweet_count": 81,
  "favorite_count": 741,
  "hashtags": [],
  "symbols": [],
  "user_mentions": [],
  "urls": [],
  "media": [
    {
      "media_url": "https://pbs.twimg.com/media/GZesbOBWQAAaLK9.jpg",
      "type": "photo"
    }
  ],
  "url": "https://twitter.com/virattt/status/1844132160195768579",
  "created_at": "2024-10-09T21:45:43.000Z",
  "#sort_index": "1844132160195768579",
  "view_count": 71992,
  "quote_count": 1,
  "is_quote_tweet": false,
  "is_retweet": false,
  "is_pinned": false,
  "is_truncated": true,
  "startUrl": "https://x.com/virattt/status/1844132160195768579"
}