Diffusers
Safetensors
nielsr HF Staff commited on
Commit
52788eb
·
verified ·
1 Parent(s): cce8309

Add model card, pipeline tag, and links to paper/code

Browse files

This PR introduces a model card for **WorldDirector-14B**, linking it to the paper [WorldDirector: Building Controllable World Simulators with Persistent Dynamic Memory](https://huggingface.co/papers/2607.02517). It also adds the `image-to-video` pipeline tag to the metadata to make the model easily discoverable on the Hugging Face Hub under the correct category, along with links to the project page and the official GitHub repository.

Files changed (1) hide show
  1. README.md +35 -0
README.md CHANGED
@@ -1,3 +1,38 @@
1
  ---
2
  license: cc-by-nc-sa-4.0
 
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  license: cc-by-nc-sa-4.0
3
+ pipeline_tag: image-to-video
4
  ---
5
+
6
+ # WorldDirector: Building Controllable World Simulators with Persistent Dynamic Memory
7
+
8
+ [📄 Paper](https://huggingface.co/papers/2607.02517) | [🌐 Project Page](https://worlddirector.github.io/) | [🐙 GitHub Repository](https://github.com/pPetrichor/WorldDirector)
9
+
10
+ **WorldDirector** is a highly controllable video world model framework designed for persistent dynamic object memory and unrestricted viewpoint exploration.
11
+
12
+ Unlike existing world models that entangle physical dynamics with pixel rendering and rely on continuous visual observation to sustain motion, this framework explicitly decouples semantic motion orchestration from visual generation. By leveraging an LLM to coordinate 3D trajectories with camera movements and subsequently employing these orchestrated trajectories as control signals for video generation, this approach ensures strict physical logic and appearance stability, successfully preserving the exact visual identities of dynamic entities even when they re-enter the scene after prolonged periods out of view.
13
+
14
+ ## Setup & Inference
15
+
16
+ For instructions on environment setup, preparing latents/conditions, and running the causal video generation pipeline, please refer to the [WorldDirector GitHub repository](https://github.com/pPetrichor/WorldDirector).
17
+
18
+ ### Model Download via CLI
19
+
20
+ You can download the model weights using `huggingface-cli` as follows:
21
+
22
+ ```shell
23
+ pip install "huggingface_hub[cli]"
24
+ huggingface-cli download hlwang06/WorldDirector --local-dir ./WorldDirector-14B
25
+ ```
26
+
27
+ ## Citation
28
+
29
+ If you find this work useful, please consider citing our paper:
30
+
31
+ ```bibtex
32
+ @article{wang2026worlddirector,
33
+ title={WorldDirector: Building Controllable World Simulators with Persistent Dynamic Memory},
34
+ author={Hanlin Wang and Hao Ouyang and Qiuyu Wang and Wen Wang and Qingyan Bai and Ka Leong Cheng and Yue Yu and Yixuan Li and Yihao Meng and Zichen Liu and Yanhong Zeng and Yujun Shen and Qifeng Chen},
35
+ journal={arXiv preprint arXiv:2607.02517},
36
+ year={2026}
37
+ }
38
+ ```