🔥 Official implementation of "VSTAR: Generative Temporal Nursing for Longer Dynamic Video Synthesis" (ICLR2025)
🚀TL;DR: VSTAR enables pretrained text-to-video models to generate longer videos with dynamic visual evolution in a single pass, without finetuning needed.
 |
 |
 |
 |
Our environment is built on top of VideoCrafter2:
conda create -n vstar python=3.10.6 pip jupyter jupyterlab matplotlib
conda activate vstar
pip install -r requirements.txt
Download pretrained Videocafter2 320x512 checkpoint from here and store it in the checkpoint folder.
Run inference_VSTAR.ipynb for testing.
This project is open-sourced under the AGPL-3.0 license. See the LICENSE file for details.
For a list of other open source components included in this project, see the file 3rd-party-licenses.txt.
This software is a research prototype, solely developed for and published as part of the publication cited above.
Please feel free to open an issue or contact personally if you have questions, need help, or need explanations. Don't hesitate to write an email to the following email address: liyumeng07@outlook.com