liqiang/ai-agent-book
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

ai-agent-book 精选快照（<2MB 代码与文档，来自 github.com/bojieli/ai-agent-book）
Build latest book artifacts / build (push) Canceled after 0s
Details
dependency resolution / resolve (3.11) (push) Canceled after 0s
Details
dependency resolution / resolve (3.13) (push) Canceled after 0s
Details
deploy-pages / build (push) Canceled after 0s
Details
deploy-pages / deploy (push) Canceled after 0s
Details
i18n consistency check / check (push) Canceled after 0s
Details
provider adoption tests / test (chapter2/context-compression) (push) Canceled after 0s
Details
provider adoption tests / test (chapter2/prompt-injection) (push) Canceled after 0s
Details
provider adoption tests / test (chapter2/system-hint) (push) Canceled after 0s
Details
provider adoption tests / test (chapter3/log-sanitization) (push) Canceled after 0s
Details
web-search-agent tests / test (push) Canceled after 0s
Details
web-search-agent tests / agentbook (push) Canceled after 0s
Details

Browse Source
This commit is contained in:
liqiang
2026-08-20 13:12:50 +00:00
commit b119135836
10275 changed files with 3284984 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
book-ja/images/fig6-9.svg
+60
View File
@@ -0,0 +1,60 @@
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 40 780 480" width="780" height="480" style="background:#ffffff">
<defs><marker id="ah" markerWidth="12" markerHeight="8" refX="12" refY="4" orient="auto"><polygon points="0 0, 12 4, 0 8" fill="#333333"/></marker><marker id="ah-light" markerWidth="12" markerHeight="8" refX="12" refY="4" orient="auto"><polygon points="0 0, 12 4, 0 8" fill="#999999"/></marker></defs>
<rect x="32" y="58" width="170" height="195" rx="6" fill="#f0f0f0" stroke="#333333" stroke-width="2"/>
<text x="117" y="80" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="16" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">OpenAI Realtime</text>
<rect x="42" y="94" width="150" height="24" rx="3" fill="#f5f5f5" stroke="#333333" stroke-width="2"/>
<text x="117" y="106" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="normal">gpt-realtime</text>
<text x="117" y="136" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="10.5" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">サーバー側 VAD + ターン検出</text>
<text x="117" y="158" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="13.5" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">非同期 function call</text>
<text x="117" y="180" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="10" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">割り込み: 現在の生成を停止</text>
<text x="117" y="200" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="8.5" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">本質的には依然として VAD の枠組み内</text>
<text x="117" y="222" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="11" fill="#999999" text-anchor="middle" dominant-baseline="central" font-weight="normal">2024年 GPT-4o ベースの preview</text>
<rect x="214" y="58" width="170" height="195" rx="6" fill="#f0f0f0" stroke="#333333" stroke-width="2"/>
<text x="299" y="80" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="16" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">Gemini Live</text>
<rect x="224" y="94" width="150" height="24" rx="3" fill="#f5f5f5" stroke="#333333" stroke-width="2"/>
<text x="299" y="106" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="normal">Gemini 2.0</text>
<text x="299" y="136" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="12.5" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">VAD 感度を設定可能</text>
<text x="299" y="158" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="10" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">割り込み時も送信済み情報を保持</text>
<text x="299" y="180" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="11" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">対話の一貫性を最適化</text>
<text x="299" y="202" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="13.5" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">OpenAI と同様の戦略</text>
<rect x="396" y="58" width="170" height="195" rx="6" fill="#d0d0d0" stroke="#333333" stroke-width="2"/>
<text x="481" y="80" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="16" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">Qwen3-Omni</text>
<rect x="406" y="94" width="150" height="24" rx="3" fill="#f5f5f5" stroke="#333333" stroke-width="2"/>
<text x="481" y="106" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="normal">Thinker-Talker MoE</text>
<text x="481" y="136" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="8.5" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">思考（理解）+ 発話（生成）</text>
<text x="481" y="158" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="11.5" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">マルチコードブックの音声符号化</text>
<text x="481" y="180" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="13.5" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">初回パケット遅延 234ms</text>
<text x="481" y="202" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="10.5" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">36件中22件のベンチマークで SOTA を達成</text>
<rect x="578" y="58" width="170" height="195" rx="6" fill="#f0f0f0" stroke="#333333" stroke-width="2"/>
<text x="663" y="80" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="16" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">Step-Audio 2</text>
<rect x="588" y="94" width="150" height="24" rx="3" fill="#f5f5f5" stroke="#333333" stroke-width="2"/>
<text x="663" y="106" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="11.5" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="normal">潜在音声符号化 + RL</text>
<text x="663" y="136" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">エンドツーエンドの音声トークン</text>
<text x="663" y="158" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="9.5" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">パラ言語理解 83.09%</text>
<text x="663" y="180" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="13" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">CoT 思考 + RAG + ツール</text>
<text x="663" y="202" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="10" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">Qwen2.5-Omni（44.18%）を上回る</text>
<rect x="40" y="273" width="700" height="100" rx="4" fill="#f5f5f5" stroke="#333333" stroke-width="2"/>
<text x="390" y="295" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="19.5" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">共通の限界: 依然としてターン検出に依存</text>
<text x="180" y="321" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="16" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">ターン検出が不十分</text>
<text x="180" y="341" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">ポーズ ≠ 発話の終わり</text>
<text x="180" y="357" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">能動的な誘導が必要</text>
<text x="400" y="321" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="16" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">割り込み判定が原始的</text>
<text x="400" y="341" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">相槌 ≠ 割り込み</text>
<text x="400" y="357" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">ノイズを発話と誤認</text>
<text x="620" y="321" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="16" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">根本原因</text>
<text x="620" y="341" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">音響信号のみ</text>
<text x="620" y="357" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">意味を捉えられない</text>
<rect x="40" y="388" width="700" height="120" rx="6" fill="#f0f0f0" stroke="#333333" stroke-width="2"/>
<text x="390" y="408" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="20" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">入出力アーキテクチャの比較</text>
<text x="60" y="430" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#333333" text-anchor="start" dominant-baseline="central" font-weight="bold">従来の直列型</text>
<text x="210" y="430" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="11" fill="#666666" text-anchor="start" dominant-baseline="central" font-weight="normal">音声→VAD→ASR→テキスト→LLM→テキスト→TTS→音声</text>
<text x="60" y="445" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#333333" text-anchor="start" dominant-baseline="central" font-weight="bold">OpenAI/Gemini</text>
<text x="210" y="445" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="11" fill="#666666" text-anchor="start" dominant-baseline="central" font-weight="normal">音声→エンドツーエンドモデル（VAD 内蔵）→音声/テキスト</text>
<text x="60" y="460" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#333333" text-anchor="start" dominant-baseline="central" font-weight="bold">Qwen3-Omni</text>
<text x="210" y="460" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="11" fill="#666666" text-anchor="start" dominant-baseline="central" font-weight="normal">音声/画像/動画→Thinker (MoE)→Talker→Speech Codec</text>
<text x="60" y="475" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#333333" text-anchor="start" dominant-baseline="central" font-weight="bold">Step-Audio 2</text>
<text x="210" y="475" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="11" fill="#666666" text-anchor="start" dominant-baseline="central" font-weight="normal">生音声→Latent Encoder→LLM+RL→音声+テキストトークン</text>
<text x="60" y="491" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#333333" text-anchor="start" dominant-baseline="central" font-weight="bold">全二重 Moshi</text>
<text x="210" y="491" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="11" fill="#666666" text-anchor="start" dominant-baseline="central" font-weight="normal">並列の2つの音声ストリーム + 内的独白テキストストリーム → VAD / ターン検出が不要</text>
</svg>
Side by Side

After

Width:  |  Height:  |  Size: 13 KiB