{"image_url":"https://cdn-ak.f.st-hatena.com/images/fotolife/s/swx-murakami/20251215/20251215223510.png","description":"\u30b5\u30fc\u30d0\u30fc\u30ef\u30fc\u30af\u30b9\u306e\u6751\u4e0a\u3067\u3059\u3002 \u3053\u306e\u30d6\u30ed\u30b0\u3067\u306f\u3001AWS re:Invent 2025 \u3067\u767a\u8868\u3055\u308c\u305f Amazon Bedrock \u306e Reinforcement Fine-tuning\uff08RFT\uff09\u306b\u3064\u3044\u3066\u7d39\u4ecb\u3057\u307e\u3059\u3002 \u524d\u63d0\uff08\u6642\u9593\u304c\u306a\u3044\u65b9\u306f\u8aad\u307f\u98db\u3070\u3057\u3066\u304f\u3060\u3055\u3044\uff09 LLM\u306e\u4e00\u822c\u7684\u306a\u5b66\u7fd2\u30d7\u30ed\u30bb\u30b9 \u4e8b\u524d\u5b66\u7fd2\uff08Pre-training\uff09 \u6559\u5e2b\u3042\u308a\u30d5\u30a1\u30a4\u30f3\u30c1\u30e5\u30fc\u30cb\u30f3\u30b0\uff08SFT / Supervised Fine-Tuning\uff09 \u30d5\u30a3\u30fc\u30c9\u30d0\u30c3\u30af\u306b\u3082\u3068\u3065\u304f\u5f37\u5316\u5b66\u7fd2\uff08RFT\uff09 RFT\u304c\u6ce8\u76ee\u3055\u308c\u308b\u80cc\u666f PPO\uff08Proximal Policy Optimization\uff09\u3068GRPO\uff08Group Relative Poli\u2026","author_url":"https://blog.hatena.ne.jp/swx-murakami/","provider_name":"Hatena Blog","categories":["Amazon Bedrock","\u6a5f\u68b0\u5b66\u7fd2"],"height":"190","html":"<iframe src=\"https://hatenablog-parts.com/embed?url=https%3A%2F%2Fblog.serverworks.co.jp%2Fbedrock-reinforcement-fine-tuning\" title=\"Amazon Bedrock \u3067\u65b0\u767b\u5834\u3057\u305fReinforcement Fine-tuning\uff08RFT\uff09\u3068\u306f - \u30b5\u30fc\u30d0\u30fc\u30ef\u30fc\u30af\u30b9\u30a8\u30f3\u30b8\u30cb\u30a2\u30d6\u30ed\u30b0\" class=\"embed-card embed-blogcard\" scrolling=\"no\" frameborder=\"0\" style=\"display: block; width: 100%; height: 190px; max-width: 500px; margin: 10px 0px;\"></iframe>","author_name":"swx-murakami","width":"100%","title":"Amazon Bedrock \u3067\u65b0\u767b\u5834\u3057\u305fReinforcement Fine-tuning\uff08RFT\uff09\u3068\u306f","blog_title":"\u30b5\u30fc\u30d0\u30fc\u30ef\u30fc\u30af\u30b9\u30a8\u30f3\u30b8\u30cb\u30a2\u30d6\u30ed\u30b0","url":"https://blog.serverworks.co.jp/bedrock-reinforcement-fine-tuning","published":"2025-12-17 06:29:03","type":"rich","provider_url":"https://hatena.blog","blog_url":"https://blog.serverworks.co.jp/","version":"1.0"}