Team Ai
Modelpublic

uva-cv-lab/FrameINO_CogVideoX_Stage2_MotionINO_v1.0

sourceHugging Facegpl-3.0updated 11mo agoView on Hugging Face
3likes75downloads
Model Card

<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/FrameINOdata target="blank"><img src=https://img.shields.io/badge/%F0%9F%A4%97%20Dataset-276cb4.svg height=22px></a> <a href=https://huggingface.co/spaces/HikariDawn/FrameINO target="blank"><img src=https://img.shields.io/badge/%F0%9F%A4%97%20OnlineDemo-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 🤗

ModelDescriptionHuggingface
CogVideoX-I2V-5B V1.0 (Stage 1 - Motion Control)Paper Weight v1.0Download
CogVideoX-I2V-5B (Stage 2 - Motion + In-N-Out Control)Paper Weight v1.0Download
Wan2.2-TI2V-5B (Stage 1 - Motion Control)New Weight v1.5 on 704PDownload
Wan2.2-TI2V-5B (Stage 2 - Motion + In-N-Out Control)New Weight v1.5 on 704PDownload
Wan2.2-TI2V-5B (Stage 2 - Motion + In-N-Out Control)New Weight v1.6 on Arbitrary ResolutionDownload

📚 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}
}