{"version":"1.0","type":"rich","published":"2023-10-22 11:39:41","html":"<iframe src=\"https://hatenablog-parts.com/embed?url=https%3A%2F%2Fwanttobejk.com%2Fentry%2F2023%2F10%2F22%2F113941\" title=\"PPO\u306b\u307e\u3064\u308f\u308b\u5099\u5fd8\u9332 - JK\u306b\u306a\u308a\u305f\u3044\" class=\"embed-card embed-blogcard\" scrolling=\"no\" frameborder=\"0\" style=\"display: block; width: 100%; height: 190px; max-width: 500px; margin: 10px 0px;\"></iframe>","blog_url":"https://wanttobejk.com/","description":"\u4f55\u306e\u8a18\u4e8b\u304b \u6700\u8fd1\u3001HuggingFace\u306e\u5f37\u5316\u5b66\u7fd2\u30c1\u30e5\u30fc\u30c8\u30ea\u30a2\u30eb\u3092\u307f\u3066\u305f\u3093\u3067\u3059\u3088\u3002 \u3067\u3001PPO\u306e\u3053\u306e\u6700\u5f8c\u306e\u76ee\u7684\u95a2\u6570\u306e\u610f\u5473\u304c\u308f\u304b\u3089\u306a\u304b\u3063\u305f\u3093\u3067\u3059\u3088\u306d\u3002 \u3053\u308c\u3002 (1) \u5404\u9805\u306f\u300c\u30af\u30ea\u30c3\u30d7\u3055\u308c\u305f\u4ee3\u7406\u65b9\u7b56\u76ee\u7684\u95a2\u6570 - \u4fa1\u5024\u95a2\u6570\u306e\u76ee\u7684\u95a2\u6570 + \u30a8\u30f3\u30c8\u30ed\u30d4\u30fc\u30dc\u30fc\u30ca\u30b9\u300d\u3068\u306a\u3063\u3066\u3044\u307e\u3059\u3002 \u672c\u8a18\u4e8b\u306f\u3053\u308c\u3092\u7406\u89e3\u3059\u308b\u305f\u3081\u306e\u5099\u5fd8\u9332\u3067\u3059\u3002\u305f\u3060\u3001\u591a\u5206\u8272\u3005\u89e3\u91c8\u9593\u9055\u3063\u3066\u308b\u3093\u3067\u3054\u6307\u6458\u3044\u305f\u3060\u3051\u308b\u3068\u5b09\u3057\u3044\u3067\u3059\u3002 \u306a\u305c1\u3064\u306e\u76ee\u7684\u95a2\u6570\u3067Actor\u3068Critic\u3092\u66f4\u65b0\u3067\u304d\u308b\u306e\uff1f \u4e00\u822c\u7684\u306aActorCritic\u3067\u306f\u3001 Actor\u306f (2) Critic\u306f (3) \u3067\u6700\u9069\u5316\u3057\u307e\u3057\u3087\u3046\u3001\u3068\u306a\u3063\u3066\u3044\u307e\u3057\u305f\u3002\u305d\u308c\u305e\u308c\u76ee\u7684\u95a2\u6570\u304c\u5b9a\u7fa9\u3055\u308c\u3001\u305d\u308c\u305e\u308c\u6700\u2026","title":"PPO\u306b\u307e\u3064\u308f\u308b\u5099\u5fd8\u9332","author_url":"https://blog.hatena.ne.jp/sakata_harumi/","provider_url":"https://hatena.blog","height":"190","author_name":"sakata_harumi","width":"100%","url":"https://wanttobejk.com/entry/2023/10/22/113941","categories":["\u6a5f\u68b0\u5b66\u7fd2"],"provider_name":"Hatena Blog","image_url":null,"blog_title":"JK\u306b\u306a\u308a\u305f\u3044"}