File size: 2,591 Bytes
8f5b101
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
cf05005
8f5b101
6c338d8
8f5b101
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
88d62a8
 
 
 
8f5b101
 
 
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
---
license: apache-2.0
base_model: Wan-AI/Wan2.1-T2V-1.3B
pipeline_tag: text-to-video
tags:
- video-generation
- autoregressive
- long-video
---

# Disen-Forcing

[Paper](https://arxiv.org/abs/2611.02822) | [Project page](https://disen-forcing.github.io/) | [Code](https://github.com/Dijkstra14/Disen-Forcing)

Disen-Forcing is an autoregressive text-to-video model distilled from Wan2.1-T2V-1.3B. It generates 832×480 videos at 16 fps from 5 s to 1 min in 4 sampling steps per chunk, keeping long-range consistency without the low-dynamic motion that a frame sink tends to cause.

## Model

- `disen_forcing.safetensors`: model weights (bf16).
- Student: Wan2.1-T2V-1.3B, initialized from the Causal Forcing chunk-wise ODE checkpoint; teacher: Wan2.1-T2V-14B.
- Training: DMD distillation with the improved recipe described in the [code repository](https://github.com/Dijkstra14/Disen-Forcing#improved-training-recipe).

## Usage

```bash
git clone https://github.com/Dijkstra14/Disen-Forcing && cd Disen-Forcing
# environment: see the repository README
huggingface-cli download Wan-AI/Wan2.1-T2V-1.3B --local-dir wan_models/Wan2.1-T2V-1.3B
huggingface-cli download ZY-Dijkstra/Disen-Forcing disen_forcing.safetensors --local-dir checkpoints

python inference.py --config configs/inference.yaml --checkpoint checkpoints/disen_forcing.safetensors \
    --prompts prompts/MovieGenVideoBench.txt --extended_prompts prompts/MovieGenVideoBench_extended.txt \
    --output_dir videos --num_output_frames 126 --seed 0
```

`--num_output_frames` counts latent frames: 21 ≈ 5 s, 126 ≈ 30 s, 252 ≈ 1 min.

## Citation

```bibtex
@inproceedings{yu2026disenforcing,
  title     = {Disen-Forcing: Disentangling Semantic Anchoring from Motion for Autoregressive Video Diffusion},
  author    = {Yu, Zhengyang and Hayakawa, Akio and Ishii, Masato and Yu, Qingtao and Li, Xuesong and Shibuya, Takashi and Mitsufuji, Yuki and Zhang, Jing},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year      = {2026}
}
```

## Acknowledgements

This model is distilled from [Wan2.1](https://github.com/Wan-Video/Wan2.1) and initialized from [Causal Forcing](https://github.com/thu-ml/Causal-Forcing). The code builds on [CausVid](https://github.com/tianweiy/CausVid), [Self-Forcing](https://github.com/guandeh17/Self-Forcing), [Infinite-Forcing](https://github.com/SOTAMak1r/Infinite-Forcing), [Causal Forcing](https://github.com/thu-ml/Causal-Forcing) and [Self Gradient Forcing](https://github.com/zhuang2002/Self_Gradient_Forcing).

## License

Apache-2.0, following Wan2.1.