--- title: FantasyWorld-Wan2.2-Fun-A14B-Control-Camera canonical_url: "https://www.modelscope.cn/models/amap_cvlab/FantasyWorld-Wan2.2-Fun-A14B-Control-Camera" md_url: "https://www.modelscope.cn/models/amap_cvlab/FantasyWorld-Wan2.2-Fun-A14B-Control-Camera.md" repository: amap_cvlab/FantasyWorld-Wan2.2-Fun-A14B-Control-Camera chinese_name: FantasyWorld-Wan2.2-Fun-A14B-Control-Camera last_updated: 2026-03-09 license: "Apache License 2.0" base_model: - PAI/Wan2.2-Fun-A14B-Control-Camera base_model_relation: finetune language: - en - zh downloads: 107 stars: 2 --- # FantasyWorld-Wan2.2-Fun-A14B-Control-Camera > FantasyWorld-Wan2.2-Fun-A14B-Control-Camera - amap_cvlab 在 ModelScope 开源的模型。FantasyWorld: Geometry-Consistent World Modeling via Unified Video and 3D Prediction amap_cvlab/FantasyWorld-Wan2.2-Fun-A14B-Control-Camera 是 ModelScope 魔搭社区上的机器学习模型，采用 Apache License 2.0 许可，基于 PAI/Wan2.2-Fun-A14B-Control-Camera 构建。 - **Repository**: amap_cvlab/FantasyWorld-Wan2.2-Fun-A14B-Control-Camera - **License**: Apache License 2.0 - **Base model**: PAI/Wan2.2-Fun-A14B-Control-Camera - **Downloads**: 107 - **Stars**: 2 - **Last updated**: 2026-03-09 Source: https://www.modelscope.cn/models/amap_cvlab/FantasyWorld-Wan2.2-Fun-A14B-Control-Camera --- # FantasyWorld: Geometry-Consistent World Modeling via Unified Video and 3D Prediction [![Home Page](https://img.shields.io/badge/🌐%20%20Project-FantasyWorld-blue.svg)](https://fantasy-amap.github.io/fantasy-world/) [![arXiv](https://img.shields.io/badge/Arxiv-2509.21657-b31b1b.svg?logo=arXiv)](https://arxiv.org/abs/2509.21657) [![Code](https://img.shields.io/badge/Code-GitHub-181717.svg?logo=GitHub)](https://github.com/Fantasy-AMAP/fantasy-world.git) [![HuggingFace](https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Wan2.1-FFD21E)](https://huggingface.co/acvlab/FantasyWorld-Wan2.1-I2V-14B-480P) [![HuggingFace](https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Wan2.2-FFD21E)](https://huggingface.co/acvlab/FantasyWorld-Wan2.2-Fun-A14B-Control-Camera) [![ModelScope](https://img.shields.io/badge/ModelScope-Wan2.1-624AFF)](https://modelscope.cn/amap_cvlab/FantasyWorld-Wan2.1-I2V-14B-480P) [![ModelScope](https://img.shields.io/badge/ModelScope-Wan2.2-624AFF)](https://modelscope.cn/amap_cvlab/FantasyWorld-Wan2.2-Fun-A14B-Control-Camera) ## 🔥🔥🔥 最新动态 - 👋 2026年2月：我们正式发布了 FantasyWorld 的代码和模型权重。 - 🏛 2026年1月：FantasyWorld 被 **ICLR 2026** 接收。 - 🎉 2025年12月：FantasyWorld 在 [WorldScore](https://huggingface.co/spaces/Howieeeee/WorldScore_Leaderboard) 排行榜（由斯坦福大学李飞飞教授团队推出）中荣获 **第一名**，在与全球最先进模型（SOTA）的对比中验证了我们方法的有效性。 ## 🌟 概览 ![Overview](https://github.com/Fantasy-AMAP/fantasy-world/blob/edd1ad8ce5eb3485c75f0b796ff74d5791a574fa/assets/overview.png?raw=true) FantasyWorld 是一个用于联合视频和 3D 场景生成的统一前馈模型。其前端采用**预调节模块 (Preconditioning Blocks, PCBs)**，通过复用冻结的 WanDiT 去噪器来提供部分去噪的隐变量，确保几何路径在有意义的特征上而非纯噪声上进行操作。骨干网络由堆叠的**重建生成一体化模块 (Integrated Reconstruction and Generation, IRG) 块**组成，在多模态条件下迭代优化视频隐变量和几何特征。每个 IRG 模块包含一个非对称的双分支结构：用于外观合成的**想象先验分支 (Imagination Prior Branch)** 和用于显式 3D 推理的**几何一致性分支 (Geometry-Consistent Branch)**，两者通过轻量级适配器和交叉注意力机制进行耦合。 ### 🚀 训练策略 FantasyWorld 利用一种**两阶段训练策略**来实现稳健的视频与 3D 的联合生成： - **阶段 1 (几何预训练)：** 利用 VGGT 风格的模型，对深度、点云和相机轨迹进行精确估计。 - **阶段 2 (联合生成)：** 一个统一的模型，无缝集成了阶段 1 的几何骨干网络与 Wan 视频生成流程。 ### 📦 模型库 我们提供两个版本的模型，以满足不同的研究和应用需求： | 模型名称 | 描述 | | :--- | :--- | | `FantasyWorld-Wan2.1-I2V-14B-480P` | **侧重复现性：** 严格遵守我们论文中详述的原始配置。最适合学术基准测试和复现报告结果。 | | `FantasyWorld-Wan2.2-Fun-A14B-Control-Camera` | **侧重性能：** 提供实质性的增强，包括升级的视频基础模型、更大规模的训练数据集以及更高的输出分辨率。 | ## 🚀 快速开始 ### 安装 1. **克隆仓库** ```bash git clone https://github.com/Fantasy-AMAP/fantasy-world.git cd fantasy-world ``` 2. **安装依赖** ```bash conda create -n fantasyworld python=3.10 conda activate fantasyworld pip install -r requirements.txt pip install thirdparty/utils3d/ ``` ### 1. FantasyWorld-Wan2.1-I2V-14B-480P #### 1.1 模型下载 | 模型 | 下载链接 | 备注 | | --------------|-------------------------------------------------------------------------------|-------------------------------| | Wan2.1-I2V-14B-480P | 🤗 [Huggingface](https://huggingface.co/Wan-AI/Wan2.1-I2V-14B-480P) 🤖 [ModelScope](https://www.modelscope.cn/models/Wan-AI/Wan2.1-I2V-14B-480P) | Base Model | FantasyWorld-Wan2.1-I2V-14B-480P | 🤗 [Huggingface](https://huggingface.co/acvlab/FantasyWorld-Wan2.1-I2V-14B-480P) 🤖 [ModelScope](https://www.modelscope.cn/models/amap_cvlab/FantasyWorld-Wan2.1-I2V-14B-480P) | FantasyWorld 使用 `huggingface` 下载模型： ```bash pip install -U "huggingface_hub" hf download "Wan-AI/Wan2.1-I2V-14B-480P" --local-dir ./models/Wan-AI/Wan2.1-I2V-14B-480P hf download "acvlab/FantasyWorld-Wan2.1-I2V-14B-480P" --local-dir ./models/FantasyWorld-Wan2.1-I2V-14B-480P/ ``` 使用 `modelscope` 下载模型： ```bash pip install -U modelscope modelscope download "Wan-AI/Wan2.1-I2V-14B-480P" --local_dir ./models/Wan-AI/Wan2.1-I2V-14B-480P modelscope download "amap_cvlab/FantasyWorld-Wan2.1-I2V-14B-480P" --local_dir ./models/FantasyWorld-Wan2.1-I2V-14B-480P/ ``` #### 1.2 推理命令 ```bash python inference_wan21.py \ --wan_ckpt_path ./models/Wan-AI/Wan2.1-I2V-14B-480P \ --model_ckpt ./models/FantasyWorld-Wan2.1-I2V-14B-480P/model.pth \ --image_path ./examples/images/input_image.png \ --camera_json_path ./examples/cameras/camera_data.json \ --prompt "In the Open Loft Living Room, sunlight streams through large windows, highlighting the sleek fireplace and elegant wooden stairs." \ --output_dir ./output-wan21 \ --sample_steps 50 \ --using_scale True ``` **参数说明:** - `--wan_ckpt_path` - **必填**: Wan模型checkpoint目录 - `--model_ckpt` - **必填**: 训练好的模型checkpoint路径 - `--image_path` - **必填**: 输入图片路径 - `--camera_json_path` - **必填**: 相机轨迹路径 - `--prompt` - **必填**: 文本提示词 - `--output_dir` - **可选**: 输出目录，默认为输入图片所在目录 - `--sample_steps` - **可选**: 采样步数，默认50 - `--using_scale` - **可选**: 是否使用scale归一化，默认True ### 2. FantasyWorld-Wan2.2-Fun-A14B-Control-Camera #### 2.1 模型下载 | Models | Download Link | Notes | | --------------|-------------------------------------------------------------------------------|-------------------------------| | Wan2.2-Fun-A14B-Control-Camera | 🤗 [Huggingface](https://huggingface.co/alibaba-pai/Wan2.2-Fun-A14B-Control-Camera) 🤖 [ModelScope](https://www.modelscope.ai/models/PAI/Wan2.2-Fun-A14B-Control-Camera) | Base Model | Wan2.2-Fun-Reward-LoRAs | 🤗 [Huggingface](https://huggingface.co/alibaba-pai/Wan2.2-Fun-Reward-LoRAs) 🤖 [ModelScope](https://www.modelscope.ai/models/PAI/Wan2.2-Fun-Reward-LoRAs) | LoRA Model | FantasyWorld-Wan2.2-Fun-A14B-Control-Camera | 🤗 [Huggingface](https://huggingface.co/acvlab/FantasyWorld-Wan2.2-Fun-A14B-Control-Camera) 🤖 [ModelScope](https://www.modelscope.ai/models/amap_cvlab/FantasyWorld-Wan2.2-Fun-A14B-Control-Camera) | FantasyWorld 使用 `huggingface` 下载模型： ```bash pip install -U "huggingface_hub" hf download "alibaba-pai/Wan2.2-Fun-A14B-Control-Camera" --local-dir ./models/PAI/Wan2.2-Fun-A14B-Control-Camera hf download "alibaba-pai/Wan2.2-Fun-Reward-LoRAs" --local-dir ./models/PAI/Wan2.2-Fun-Reward-LoRAs hf download "acvlab/FantasyWorld-Wan2.2-Fun-A14B-Control-Camera" --local-dir ./models/FantasyWorld-Wan2.2-Fun-A14B-Control-Camera/ ``` 使用 `modelscope` 下载模型： ```bash pip install -U modelscope modelscope download "PAI/Wan2.2-Fun-A14B-Control-Camera" --local_dir ./models/PAI/Wan2.2-Fun-A14B-Control-Camera modelscope download "PAI/Wan2.2-Fun-Reward-LoRAs" --local_dir ./models/PAI/Wan2.2-Fun-Reward-LoRAs modelscope download "amap_cvlab/FantasyWorld-Wan2.2-Fun-A14B-Control-Camera" --local_dir ./models/FantasyWorld-Wan2.2-Fun-A14B-Control-Camera/ ``` #### 2.2 推理命令 ```bash python inference_wan22.py \ --image_path ./examples/images/input_image.png \ --end_image_path ./examples/images/end_image.png \ --wan_ckpt_path ./models/ \ --camera_json_path ./examples/cameras/camera_data.json \ --prompt "In the Open Loft Living Room, sunlight streams through large windows, highlighting the sleek fireplace and elegant wooden stairs." \ --model_ckpt_high ./models/FantasyWorld-Wan2.2-Fun-A14B-Control-Camera/high_noise_model.pth \ --model_ckpt_low ./models/FantasyWorld-Wan2.2-Fun-A14B-Control-Camera/low_noise_model.pth \ --output_dir ./output-wan22 \ --sample_steps 50 \ --using_scale True ``` **参数说明:** - `--image_path` - **必填**: 首帧图像路径 - `--end_image_path` - **必填**: 尾帧图像路径 - `--wan_ckpt_path` - **必填**: Wan模型checkpoint目录 - `--camera_json_path` - **必填**: 相机轨迹路径 - `--prompt` - **必填**: 文本提示词 - `--output_dir` - **可选**: 输出目录 - `--sample_steps` - **可选**: 采样步数，默认50 - `--using_scale` - **可选**: 是否使用scale归一化，默认True ## 🧩 社区贡献 我们无比欢迎来自开源社区的贡献！❤️ 如果您基于 FantasyWorld 进行了改进或开发了衍生项目，请务必告知我们。您也可以直接发送邮件至 [frank.jf@alibaba-inc.com](mailto://frank.jf@alibaba-inc.com)。 我们非常乐意在本项目中收录并展示您的工作，方便社区查阅、参考。 ## 🔗Citation 如果 FantasyWorld 帮助到了您，还请给我们一个 star⭐ 或引用我们的论文以示鼓励。 ``` @inproceedings{ dai2025fantasyworld, title={FantasyWorld: Geometry-Consistent World Modeling via Unified Video and 3D Prediction}, author={Yixiang Dai and Fan Jiang and Chiyu Wang and Mu Xu and Yonggang Qi}, booktitle={The Fourteenth International Conference on Learning Representations}, year={2026}, url={https://openreview.net/forum?id=3q9vHEqsNx} } ``` ## 🙏 致谢 我们非常感谢 [Wan](https://github.com/Wan-Video)、[VideoX-Fun](https://github.com/aigc-apps/VideoX-Fun)、[DiffSynth-Studio](https://github.com/modelscope/DiffSynth-Studio) 和 [VGGT](https://github.com/facebookresearch/vggt.git) 的出色工作。