Search papers, labs, and topics across Lattice.
The paper introduces RefVideo-6M, a large-scale dataset designed to improve instructional video editing by addressing the limitations of existing datasets that rely on automatic editing models and textual instructions. By utilizing artifact-free real videos as editing targets and incorporating diverse visual references, the dataset enhances the quality of supervision for training video editing models. Experimental results show that models trained on RefVideo-6M achieve superior visual quality, controllability, and reference consistency compared to those trained on traditional datasets.
RefVideo-6M revolutionizes video editing datasets by providing 5 million high-quality editing samples that prioritize visual references over flawed automatic edits.
Recent advances in video editing have been largely driven by large-scale instruction-based datasets. However, existing datasets still suffer from two critical limitations. First, target videos are commonly produced by automatic editing models, which may introduce visible artifacts and unreliable supervision signals. Second, most public datasets rely primarily on textual instructions, while lacking visual references that are crucial for precise, identity-preserving, and controllable editing. To address these limitations, we introduce RefVideo-6M, a large-scale reference-guided editing dataset containing 5 million video editing samples and 1 million image editing samples. To ensure reliable supervision, our dataset uses a construction pipeline that treats artifact-free real videos as editing targets and generates quality-filtered input conditions with multiple editing experts. In addition, it provides approximately 6 million visual references, covering diverse reference types and editing scenarios, thereby enabling models to learn fine-grained visual correspondence beyond text-only instructions. Based on RefVideo-6M, we further train a reference-guided video editing model, Ref-MoT, to evaluate the effectiveness and scalability of the proposed dataset. Extensive experiments demonstrate that RefVideo-6M provides substantially more reliable supervision than existing datasets and enables the training of powerful editing models with improved visual quality, controllability, and reference consistency. The open-source dataset is available at https://huggingface.co/datasets/RefVideo6M/RefVideo6M.
