File size: 2,708 Bytes
3c8759a
e244891
 
3c8759a
 
 
 
 
 
e244891
3c8759a
 
3d20648
 
 
 
 
 
 
e244891
3d20648
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e244891
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
---
language:
- en
pipeline_tag: image-to-video
tags:
- image-to-video
- audio-conditioned
- diffusion
- talking-avatar
- pytorch
---

<div align="center">

<h1>AvatarForcing</h1>

<h3>One-Step Streaming Talking Avatars via Local-Future Sliding-Window Denoising</h3>

<p>
  <a href="https://huggingface.co/papers/2603.14331"><img src="https://img.shields.io/badge/HuggingFace-Paper-ffbd45?style=for-the-badge&logo=huggingface&logoColor=white" alt="Hugging Face Paper"></a>
  <a href="https://arxiv.org/abs/2603.14331"><img src="https://img.shields.io/badge/arXiv-2603.14331-b31b1b?style=for-the-badge" alt="arXiv"></a>
  <a href="https://cuiliyuan121.github.io/AvatarForcing/"><img src="https://img.shields.io/badge/Project-Page-blue?style=for-the-badge&logo=googlechrome&logoColor=white" alt="Project Page"></a>
</p>

</div>

AvatarForcing is a **one-step streaming diffusion** framework for talking avatars. It generates video from **one reference image + speech audio + (optional) text prompt**, using **local-future sliding-window denoising** with **heterogeneous noise levels** and **dual-anchor temporal forcing** for long-form stability. For method details, see: https://arxiv.org/abs/2603.14331

This Hugging Face repo (`lycui/AvatarForcing`) provides two training-stage checkpoints:

- `ode_audio_init.pt`: stage-1 **ODE** initialization weights
- `model.pt`: stage-2 **DMD** weights

## Model Download

| Models | Download Link | Notes |
|---|---|---|
| Wan2.1-T2V-1.3B | 🤗 [Huggingface](https://huggingface.co/Wan-AI/Wan2.1-T2V-1.3B) | Base model (student) |
| AvatarForcing | 🤗 [Huggingface](https://huggingface.co/lycui/AvatarForcing) | `ode_audio_init.pt` (ODE) + `model.pt` (DMD) |
| Wav2Vec | 🤗 [Huggingface](https://huggingface.co/facebook/wav2vec2-base-960h) | Audio encoder |

Download models using `huggingface-cli`:

```sh
pip install "huggingface_hub[cli]"
mkdir -p pretrained_models

huggingface-cli download Wan-AI/Wan2.1-T2V-1.3B --local-dir ./pretrained_models/Wan2.1-T2V-1.3B
huggingface-cli download facebook/wav2vec2-base-960h --local-dir ./pretrained_models/wav2vec2-base-960h
huggingface-cli download lycui/AvatarForcing --local-dir ./pretrained_models/AvatarForcing
```

<details>
<summary><strong>Citation</strong></summary>

```bibtex
@misc{cui2026avatarforcingonestepstreamingtalking,
      title={AvatarForcing: One-Step Streaming Talking Avatars via Local-Future Sliding-Window Denoising},
      author={Liyuan Cui and Wentao Hu and Wenyuan Zhang and Zesong Yang and Fan Shi and Xiaoqiang Liu},
      year={2026},
      eprint={2603.14331},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2603.14331},
}
```

</details>