V2VBench is the dataset used in the research paper Diffusion Model-Based Video Editing: A Survey.
V2VBench is designed to benchmark and evaluate diffusion model-based video editing techniques and provides a diverse collection of video samples and corresponding editing tasks, enabling researchers and practitioners to assess the performance of various video editing algorithms.
For detailed instructions on how to use the V2VBench dataset, please refer to our documentation.
