{"type":"rich","version":1.0,"provider_name":"Speaker Deck","provider_url":"https://speakerdeck.com/","title":"タスクの複雑さでモデルを選ぶ ── Thompson Samplingで動かす“トークン/コスト最適化","author_name":"satoh-y-0323","author_url":"https://speakerdeck.com/satohy0323","html":"\u003ciframe id=\"talk_frame_1566912\" class=\"speakerdeck-iframe\" src=\"https://speakerdeck.com/player/dedf0326ac8d4a8e8c666a4985bec90e\" width=\"710\" height=\"399\" style=\"aspect-ratio:710/399; border:0; padding:0; margin:0; background:transparent;\" frameborder=\"0\" allowtransparency=\"true\" allowfullscreen=\"allowfullscreen\"\u003e\u003c/iframe\u003e\n","width":710,"height":399,"ratio":1.7777777777777777}