README.md

---
license: gpl-3.0
---


<div align="center">

# Frame In-N-Out: Unbounded Controllable Image-to-Video Generation 

</div>


<div align="center">
  <a href=https://uva-computer-vision-lab.github.io/Frame-In-N-Out/ target="_blank"><img src=https://img.shields.io/badge/Project%20Page-333399.svg?logo=homepage height=22px></a>
  <a href=https://huggingface.co/collections/uva-cv-lab/frame-in-n-out target="_blank"><img src=https://img.shields.io/badge/%F0%9F%A4%97%20Models-d96902.svg height=22px></a>
  <a href=https://huggingface.co/datasets/uva-cv-lab/FrameINO_data  target="_blank"><img src=https://img.shields.io/badge/%F0%9F%A4%97%20Dataset-276cb4.svg height=22px></a>
  <a href=https://github.com/UVA-Computer-Vision-Lab/FrameINO target="_blank"><img src= https://img.shields.io/badge/Code-black.svg?logo=github height=22px></a>
  <a href=https://arxiv.org/abs/2505.21491 target="_blank"><img src=https://img.shields.io/badge/Arxiv-b5212f.svg?logo=arxiv height=22px></a>
</div>

<br>



## Intro Video
<p align="center">
  <video
    controls
    autoplay
    playsinline
    muted
    loop
    src="https://github.com/user-attachments/assets/0fabd2a4-9d3b-4148-bc04-6fc03c53caca"
    width="60%"
  >
  </video>
  <br>
  <em> Frame In-N-Out is a controllable Image-to-Video generation Diffusion Transformer model where objects can enter or exit the scene along user-specified motion trajectories and ID reference. Our method introduces a new dataset curation pattern recognition, evaluation protocol, and a <b>motion-controllable</b>, <b>identity-preserving</b>, <b>unbounded canvas</b> Video Diffusion Transformer, to achieve Frame In and Frame Out in the cinematic domain. </em>
</p>






## Model Zoo 🤗
| Model                                                          | Description                    | Huggingface                                                                                     |
|--------------------------------------------------------------- | -------------------------------| ------------------------------------------------------------------------------------------------| 
| CogVideoX-I2V-5B V1.0 (Stage 1 - Motion Control)               |  Paper Weight v1.0             |     [Download](https://huggingface.co/uva-cv-lab/FrameINO_CogVideoX_Stage1_Motion_v1.0)         |
| CogVideoX-I2V-5B  (Stage 2 - Motion + In-N-Out Control)        |  Paper Weight v1.0             |     [Download](https://huggingface.co/uva-cv-lab/FrameINO_CogVideoX_Stage2_MotionINO_v1.0)      |
| Wan2.2-TI2V-5B  (Stage 1 - Motion Control)                     |  New Weight v1.5 on 704P       |     [Download](https://huggingface.co/uva-cv-lab/FrameINO_Wan2.2_5B_Stage1_Motion_v1.5)         |
| Wan2.2-TI2V-5B  (Stage 2 - Motion + In-N-Out Control)          |  New Weight v1.5 on 704P       |     [Download](https://huggingface.co/uva-cv-lab/FrameINO_Wan2.2_5B_Stage2_MotionINO_v1.5)      |



## 📚 Citation
```bibtex
@article{wang2025frame,
  title={Frame In-N-Out: Unbounded Controllable Image-to-Video Generation},
  author={Wang, Boyang and Chen, Xuweiyi and Gadelha, Matheus and Cheng, Zezhou},
  journal={arXiv preprint arXiv:2505.21491},
  year={2025}
}
```