liqiang
b119135836
Build latest book artifacts / build (push) Canceled after 0s
dependency resolution / resolve (3.11) (push) Canceled after 0s
dependency resolution / resolve (3.13) (push) Canceled after 0s
deploy-pages / build (push) Canceled after 0s
deploy-pages / deploy (push) Canceled after 0s
i18n consistency check / check (push) Canceled after 0s
provider adoption tests / test (chapter2/context-compression) (push) Canceled after 0s
provider adoption tests / test (chapter2/prompt-injection) (push) Canceled after 0s
provider adoption tests / test (chapter2/system-hint) (push) Canceled after 0s
provider adoption tests / test (chapter3/log-sanitization) (push) Canceled after 0s
web-search-agent tests / test (push) Canceled after 0s
web-search-agent tests / agentbook (push) Canceled after 0s
67 lines
13 KiB
XML
67 lines
13 KiB
XML
<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><marker id="ah-sm" markerWidth="6" markerHeight="4" refX="6" refY="2" orient="auto"><polygon points="0 0, 6 2, 0 4" fill="#999999"/></marker></defs>
|
||
|
||
<rect x="55" y="55" width="210" height="50" rx="6" fill="#d0d0d0" stroke="#333333" stroke-width="2"/>
|
||
<text x="160" y="73" 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">RT-2</text>
|
||
<text x="160" y="93" 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"></text>
|
||
<rect x="63" y="115" width="194" height="50" rx="3" fill="#d0d0d0" stroke="#333333" stroke-width="2"/>
|
||
<text x="160" y="133" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="15" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">視覚言語バックボーン</text>
|
||
<text x="160" y="151" 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">大規模事前学習 VLM</text>
|
||
<rect x="63" y="173" width="194" height="50" rx="3" fill="#f0f0f0" stroke="#333333" stroke-width="2"/>
|
||
<text x="160" y="191" 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">行動の離散化</text>
|
||
<text x="160" y="209" 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">連続的な行動 → トークン</text>
|
||
<line x1="160" y1="165" x2="160" y2="173" stroke="#999999" stroke-width="2" marker-end="url(#ah-sm)"/>
|
||
<rect x="63" y="231" width="194" height="50" rx="3" fill="#f0f0f0" stroke="#333333" stroke-width="2"/>
|
||
<text x="160" y="249" 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">自己回帰的な出力</text>
|
||
<text x="160" y="267" 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">トークン単位で生成</text>
|
||
<line x1="160" y1="223" x2="160" y2="231" stroke="#999999" stroke-width="2" marker-end="url(#ah-sm)"/>
|
||
<text x="160" y="294" 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">VLM を直接ファインチューニング</text>
|
||
<text x="160" y="312" 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">事前学習の汎化性能を活用</text>
|
||
<text x="160" y="330" 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">新規の物体/指示へゼロショット転移</text>
|
||
<rect x="290" y="55" width="210" height="50" rx="6" fill="#d0d0d0" stroke="#333333" stroke-width="2"/>
|
||
<text x="395" y="73" 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">OpenVLA</text>
|
||
<text x="395" y="93" 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"></text>
|
||
<rect x="298" y="115" width="194" height="50" rx="3" fill="#d0d0d0" stroke="#333333" stroke-width="2"/>
|
||
<text x="395" y="133" 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">視覚エンコーダ</text>
|
||
<text x="395" y="151" 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">DINOv2 + SigLIP</text>
|
||
<rect x="298" y="173" width="194" height="50" rx="3" fill="#f0f0f0" stroke="#333333" stroke-width="2"/>
|
||
<text x="395" y="191" 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">言語モデル</text>
|
||
<text x="395" y="209" 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">Llama 2 7B</text>
|
||
<line x1="395" y1="165" x2="395" y2="173" stroke="#999999" stroke-width="2" marker-end="url(#ah-sm)"/>
|
||
<rect x="298" y="231" width="194" height="50" rx="3" fill="#f0f0f0" stroke="#333333" stroke-width="2"/>
|
||
<text x="395" y="249" 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">アクションヘッド</text>
|
||
<text x="395" y="267" 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">離散的な行動トークン</text>
|
||
<line x1="395" y1="223" x2="395" y2="231" stroke="#999999" stroke-width="2" marker-end="url(#ah-sm)"/>
|
||
<text x="395" y="294" 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">完全オープンソース · コミュニティで再現可能</text>
|
||
<text x="395" y="312" 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">Open X-Embodiment で事前学習</text>
|
||
<text x="395" y="330" 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">オリジナルは単一ステップの自己回帰 ~6Hz</text>
|
||
<rect x="525" y="55" width="210" height="50" rx="6" fill="#d0d0d0" stroke="#333333" stroke-width="2"/>
|
||
<text x="630" y="73" 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">π₀</text>
|
||
<text x="630" y="93" 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"></text>
|
||
<rect x="533" y="115" width="194" height="50" rx="3" fill="#d0d0d0" stroke="#333333" stroke-width="2"/>
|
||
<text x="630" y="133" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="15" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">視覚言語バックボーン</text>
|
||
<text x="630" y="151" 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">PaliGemma</text>
|
||
<rect x="533" y="173" width="194" height="50" rx="3" fill="#f0f0f0" stroke="#333333" stroke-width="2"/>
|
||
<text x="630" y="191" 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">アクションエキスパート</text>
|
||
<text x="630" y="209" 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">flow matching デノイジング</text>
|
||
<line x1="630" y1="165" x2="630" y2="173" stroke="#999999" stroke-width="2" marker-end="url(#ah-sm)"/>
|
||
<rect x="533" y="231" width="194" height="50" rx="3" fill="#f0f0f0" stroke="#333333" stroke-width="2"/>
|
||
<text x="630" y="249" 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">行動の出力</text>
|
||
<text x="630" y="267" 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">連続的な行動軌跡</text>
|
||
<line x1="630" y1="223" x2="630" y2="231" stroke="#999999" stroke-width="2" marker-end="url(#ah-sm)"/>
|
||
<text x="630" y="294" 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">連続軌跡生成の経路</text>
|
||
<text x="630" y="312" 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">アクションチャンキング 25〜50/チャンク・50Hz</text>
|
||
<text x="630" y="330" 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">巧緻な操作の滑らかさが向上</text>
|
||
<rect x="35" y="365" width="710" height="135" rx="4" fill="#f5f5f5" stroke="#333333" stroke-width="2"/>
|
||
<text x="390" y="387" 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">2層の制御アーキテクチャ</text>
|
||
<rect x="55" y="407" width="300" height="40" rx="6" fill="#f0f0f0" stroke="#333333" stroke-width="2"/>
|
||
<text x="205" y="420" 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">長期プランニング（0.1〜1Hz）</text>
|
||
<text x="205" y="437" 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">汎用 VLM / Gemini Robotics-ER 1.5</text>
|
||
<line x1="357" y1="427" x2="400" y2="427" stroke="#333333" stroke-width="2" marker-end="url(#ah)"/>
|
||
<text x="378.5" y="420" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="12" fill="#666666" text-anchor="middle">サブゴール</text>
|
||
<rect x="402" y="407" width="300" height="40" rx="6" fill="#d0d0d0" stroke="#333333" stroke-width="2"/>
|
||
<text x="552" y="420" 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">VLA 制御（1〜10Hz）</text>
|
||
<text x="552" y="437" 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">RT-2 / OpenVLA / π₀</text>
|
||
<text x="390" y="465" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="12.5" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">共通基盤: 模倣学習 + 大規模なクロスエンボディメントのデモデータで、データのボトルネックを打破</text>
|
||
<text x="390" y="485" 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">2つの行動表現の経路: 離散的な行動トークン（RT-2 / OpenVLA）· 連続軌跡生成（π₀）</text>
|
||
</svg> |
