-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathindex.zh.html
More file actions
172 lines (144 loc) · 16.7 KB
/
index.zh.html
File metadata and controls
172 lines (144 loc) · 16.7 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
<!DOCTYPE html>
<html lang="zh">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Fantasy AIGC Family - 以生成式 AI 统一“人-世界-交互”的新一代范式</title>
<link rel="preconnect" href="https://fonts.googleapis.com">
<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
<link href="https://fonts.googleapis.com/css2?family=Inter:wght@300;400;600&family=Space+Grotesk:wght@300;700&display=swap" rel="stylesheet">
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.4.0/css/all.min.css">
<link rel="stylesheet" href="style.css">
</head>
<body>
<div class="ambient-light one"></div>
<div class="ambient-light two"></div>
<div class="grid-overlay"></div>
<nav>
<div class="logo">Fantasy AIGC Family</div>
<div class="nav-links">
<a href="#overview">首页</a>
<a href="https://github.com/Fantasy-AMAP/">系列工作</a>
<a href="#news">动态</a>
<a href="index.html" class="lang-btn">English</a>
</div>
</nav>
<header class="hero">
<h1>Fantasy AIGC Family</h1>
<p>以生成式 AI 统一“人-世界-交互”的新一代范式</p>
</header>
<main class="container">
<!-- overview Section -->
<section id="overview" class="overview">
<h2 class="section-title"><i class="fas fa-lightbulb"></i> 总体架构</h2>
<div class="text-content">
<p>Fantasy AIGC Family 以<a href='https://www.wan-ai.co/'>通义万相</a>视频生成模型为统一基座，在同一个强视频生成内核上向空间智能、世界模型、具身智能与 AI 创意等技术领域辐射，构建覆盖<strong>人</strong>、<strong>世界</strong>、<strong>交互</strong>三大维度的三位一体系列模型与系统接口：在人维度沉淀可控表达与一致性表征以支撑可信数字人，在世界维度沉淀可探索的场景表征与可验证一致性以支撑可用世界模型，在交互维度沉淀以 Action 驱动的闭环控制与协同机制以支撑可扩展的交互系统，使三类能力共享表示、共享数据与共享工程管线，并形成可回流、可持续进化的能力飞轮。</p>
</div>
</section>
<!-- News Section -->
<section id="news" class="news-section">
<h2 class="section-title"><i class="fas fa-bolt"></i> 最新动态与里程碑</h2>
<div class="news-list">
<div class="news-item"><div class="news-text">📢 <strong>2026年2月</strong> – <a href='https://fantasy-amap.github.io/fantasy-vln/'><strong>FantasyVLN</strong></a> 正式被 CVPR 2026 收录</div></div><div class="news-item"><div class="news-text">📢 <strong>2026年1月</strong> – 我们正式开源了 <a href='https://fantasy-amap.github.io/fantasy-vln/'><strong>FantasyVLN</strong></a> 的训练和推理代码及模型权重。</div></div><div class="news-item"><div class="news-text">🏆 <strong>2025年12月</strong> - <a href='https://fantasy-amap.github.io/fantasy-world/'><strong>FantasyWorld</strong></a> 在<strong>斯坦福李飞飞教授团队</strong>发布的 <a href='https://huggingface.co/spaces/Howieeeee/WorldScore_Leaderboard'><strong>WorldScore</strong></a> 评测中斩获<strong>第一名</strong>，验证了我们的技术路线。</div></div><div class="news-item"><div class="news-text">🏛 <strong>2025年11月</strong> – 我们的两篇工作 <a href='https://fantasy-amap.github.io/fantasy-talking2/'><strong>FantasyTalking2</strong></a> 和 <a href='https://fantasy-amap.github.io/fantasy-hsi/'><strong>FantasyHSI</strong></a> 被 <strong>AAAI 2026</strong> 接收。</div></div><div class="news-item"><div class="news-text">🏛 <strong>2025年7月</strong> – <a href='https://fantasy-amap.github.io/fantasy-talking/'><strong>FantasyTalking</strong></a> 被 <strong>ACM MM 2025</strong> 接收。</div></div><div class="news-item"><div class="news-text">📢 <strong>2025年4月</strong> – 我们正式开源了 <a href='https://fantasy-amap.github.io/fantasy-talking/'><strong>FantasyTalking</strong></a> 和 <a href='https://fantasy-amap.github.io/fantasy-id/'><strong>FantasyID</strong></a> 的推理代码及模型权重。</div></div>
</div>
</section>
<!-- Projects Section -->
<section id="projects">
<article class="project-card">
<div class="video-container">
<video controls style="width: 100%; height: 100%;">
<source src="https://fantasy-amap.github.io/fantasy-vln/assets/FantasyVLN-demo.mp4" type="video/mp4">
Your browser does not support the video tag.
</video>
</div>
<div class="card-content">
<div class="project-header">
<div class="project-title">FantasyVLN</div>
<div class="project-badges"><a href="https://fantasy-amap.github.io/fantasy-vln/" target="_blank" class="badge badge-project"><i class="fas fa-globe"></i> Project</a> <a href="https://arxiv.org/abs/2601.13976" target="_blank" class="badge badge-arxiv"><i class="fas fa-file-pdf"></i> arXiv</a> <a href="https://github.com/Fantasy-AMAP/fantasy-vln" target="_blank" class="badge badge-github" data-github-repo="Fantasy-AMAP/fantasy-vln"><i class="fab fa-github"></i> Code<span class="star-suffix" style="display:none"> (<span class="star-count"></span> <i class="fas fa-star"></i>)</span></a> <a href="https://huggingface.co/acvlab/FantasyVLN" target="_blank" class="badge badge-huggingface"><i class="fas fa-face-smile"></i> HuggingFace</a> <a href="https://modelscope.cn/models/amap_cvlab/FantasyVLN" target="_blank" class="badge badge-modelscope"><i class="fas fa-cube"></i> ModelScope</a> </div>
</div>
<div class="project-desc">一个统一的多模态链式思维推理框架，通过将世界模型的推演能力内化到 VLN 架构中，基于自然语言指令和视觉观察，实现高效且精确的导航。</div>
</div>
</article>
<article class="project-card">
<div class="video-container">
<iframe width="922" height="519" src="https://www.youtube.com/embed/_gLKZL5ytEo" title="FantasyWorld" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe>
</div>
<div class="card-content">
<div class="project-header">
<div class="project-title">FantasyWorld</div>
<div class="project-badges"><a href="" target="_blank" class="badge badge-publish"><i class="fas fa-landmark"></i> ICLR 2026</a> <a href="https://fantasy-amap.github.io/fantasy-world/" target="_blank" class="badge badge-project"><i class="fas fa-globe"></i> Project</a> <a href="https://arxiv.org/abs/2509.21657" target="_blank" class="badge badge-arxiv"><i class="fas fa-file-pdf"></i> arXiv</a> <a href="https://github.com/Fantasy-AMAP/fantasy-world" target="_blank" class="badge badge-github" data-github-repo="Fantasy-AMAP/fantasy-world"><i class="fab fa-github"></i> Code<span class="star-suffix" style="display:none"> (<span class="star-count"></span> <i class="fas fa-star"></i>)</span></a> </div>
</div>
<div class="project-desc">对应"世界（Worlds）"维度。一个统一的世界模型，集成了视频先验和几何基础，用于合成可探索且几何一致的3D场景。它强调Action驱动下的时空一致性，并作为空间智能与几何一致性的可验证结构锚点。</div>
</div>
</article>
<article class="project-card">
<div class="video-container">
<iframe width="922" height="519" src="https://www.youtube.com/embed/L8iBGuuJKlE" title="FantasyTalking" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe>
</div>
<div class="card-content">
<div class="project-header">
<div class="project-title">FantasyTalking</div>
<div class="project-badges"><a href="https://dl.acm.org/doi/10.1145/3746027.3755217" target="_blank" class="badge badge-publish"><i class="fas fa-landmark"></i> ACM MM 2025</a> <a href="https://fantasy-amap.github.io/fantasy-talking/" target="_blank" class="badge badge-project"><i class="fas fa-globe"></i> Project</a> <a href="https://arxiv.org/abs/2504.04842" target="_blank" class="badge badge-arxiv"><i class="fas fa-file-pdf"></i> arXiv</a> <a href="https://github.com/Fantasy-AMAP/fantasy-talking" target="_blank" class="badge badge-github" data-github-repo="Fantasy-AMAP/fantasy-talking"><i class="fab fa-github"></i> Code<span class="star-suffix" style="display:none"> (<span class="star-count"></span> <i class="fas fa-star"></i>)</span></a> <a href="https://huggingface.co/acvlab/FantasyTalking" target="_blank" class="badge badge-huggingface"><i class="fas fa-face-smile"></i> HuggingFace</a> <a href="https://huggingface.co/spaces/acvlab/FantasyTalking" target="_blank" class="badge badge-huggingface"><i class="fas fa-face-smile"></i> HuggingFace Space</a> <a href="https://modelscope.cn/models/amap_cvlab/FantasyTalking" target="_blank" class="badge badge-modelscope"><i class="fas fa-cube"></i> ModelScope</a> </div>
</div>
<div class="project-desc">首个基于 Wan 的高保真音频驱动虚拟人系统，通过双阶段音视对齐与可控运动调制，实现动态场景下面部表情、唇动与身体姿态的精准同步。</div>
</div>
</article>
<article class="project-card">
<div class="video-container">
<iframe width="922" height="519" src="https://www.youtube.com/embed/XJAP9SsXdPg" title="FantasyTalking2" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe>
</div>
<div class="card-content">
<div class="project-header">
<div class="project-title">FantasyTalking2</div>
<div class="project-badges"><a href="https://doi.org/10.48550/arXiv.2508.11255" target="_blank" class="badge badge-publish"><i class="fas fa-landmark"></i> AAAI 2026</a> <a href="https://fantasy-amap.github.io/fantasy-talking2/" target="_blank" class="badge badge-project"><i class="fas fa-globe"></i> Project</a> <a href="https://arxiv.org/abs/2508.11255v1" target="_blank" class="badge badge-arxiv"><i class="fas fa-file-pdf"></i> arXiv</a> <a href="https://github.com/Fantasy-AMAP/fantasy-talking2" target="_blank" class="badge badge-github" ><i class="fab fa-github"></i> Coming Soon</a> </div>
</div>
<div class="project-desc">一种新颖的“时间步-网络层”自适应多专家偏好优化(TLPO)方法，在口型一致、动作自然、视觉效果三个维度上提升了音频驱动数字人动画的质量。</div>
</div>
</article>
<article class="project-card">
<div class="video-container">
<iframe width="922" height="519" src="https://www.youtube.com/embed/cgbjvqiv0sw" title="FantasyPortrait" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe>
</div>
<div class="card-content">
<div class="project-header">
<div class="project-title">FantasyPortrait</div>
<div class="project-badges"><a href="https://fantasy-amap.github.io/fantasy-portrait/" target="_blank" class="badge badge-project"><i class="fas fa-globe"></i> Project</a> <a href="https://arxiv.org/abs/2507.12956" target="_blank" class="badge badge-arxiv"><i class="fas fa-file-pdf"></i> arXiv</a> <a href="https://github.com/Fantasy-AMAP/fantasy-portrait" target="_blank" class="badge badge-github" data-github-repo="Fantasy-AMAP/fantasy-portrait"><i class="fab fa-github"></i> Code<span class="star-suffix" style="display:none"> (<span class="star-count"></span> <i class="fas fa-star"></i>)</span></a> </div>
</div>
<div class="project-desc">一种全新的表情驱动视频生成方法，将情绪增强学习与掩码交叉注意力相结合，可在单人或多人肖像场景中生成高质量且富有表现力的动画。</div>
</div>
</article>
<article class="project-card">
<div class="video-container">
<video controls style="width: 100%; height: 100%;">
<source src="https://fantasy-amap.github.io/fantasy-hsi/assets/first_video_audio.mp4" type="video/mp4">
Your browser does not support the video tag.
</video>
</div>
<div class="card-content">
<div class="project-header">
<div class="project-title">FantasyHSI</div>
<div class="project-badges"><a href="https://doi.org/10.48550/arXiv.2509.01232" target="_blank" class="badge badge-publish"><i class="fas fa-landmark"></i> AAAI 2026</a> <a href="https://fantasy-amap.github.io/fantasy-hsi/" target="_blank" class="badge badge-project"><i class="fas fa-globe"></i> Project</a> <a href="https://arxiv.org/abs/2509.01232" target="_blank" class="badge badge-arxiv"><i class="fas fa-file-pdf"></i> arXiv</a> <a href="https://github.com/Fantasy-AMAP/fantasy-hsi" target="_blank" class="badge badge-github" ><i class="fab fa-github"></i> Coming Soon</a> </div>
</div>
<div class="project-desc">面向"人与世界交互"的一种基于图结构的多智能体框架，将视频生成与三维世界动态相融合。它面向更广义的交互闭环与动作空间统一，使视频生成从内容终点转为交互系统的控制通道。</div>
</div>
</article>
<article class="project-card">
<div class="video-container">
<iframe width="922" height="519" src="https://www.youtube.com/embed/_7Mo6TrXTZs" title="FantasyID" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe>
</div>
<div class="card-content">
<div class="project-header">
<div class="project-title">FantasyID</div>
<div class="project-badges"><a href="https://fantasy-amap.github.io/fantasy-id/" target="_blank" class="badge badge-project"><i class="fas fa-globe"></i> Project</a> <a href="https://arxiv.org/abs/2502.13995" target="_blank" class="badge badge-arxiv"><i class="fas fa-file-pdf"></i> arXiv</a> <a href="https://github.com/Fantasy-AMAP/fantasy-id" target="_blank" class="badge badge-github" data-github-repo="Fantasy-AMAP/fantasy-id"><i class="fab fa-github"></i> Code<span class="star-suffix" style="display:none"> (<span class="star-count"></span> <i class="fas fa-star"></i>)</span></a> <a href="https://huggingface.co/acvlab/FantasyID" target="_blank" class="badge badge-huggingface"><i class="fas fa-face-smile"></i> HuggingFace</a> <a href="https://modelscope.cn/models/amap_cvlab/FantasyID" target="_blank" class="badge badge-modelscope"><i class="fas fa-cube"></i> ModelScope</a> </div>
</div>
<div class="project-desc">以3D面部先验、多视角增强以及层感知注入的提升运动场景下的ID保持视频生成框架。</div>
</div>
</article>
</section>
</main>
<footer>
<p>© 2026 Fantasy AIGC Team</p>
</footer>
<script src="script.js"></script>
</body>
</html>