CI-VID：一个连贯的交错文本-视频数据集
CI-VID: A Coherent Interleaved Text-Video Dataset
CI-VID 提供了带交错文本标注的多片段视频序列，可直接用于训练和评测文本-视频到视频（T&V2V）生成模型，推动从孤立片段生成向具有跨片段语义连贯性的故事化视频生成发展。
原文依据 CI-VID contains over 340,000 samples, each comprising a semantically coherent video sequence with interleaved text captions
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论文本到视频（T2V）生成近期引起了广泛关注，由此催生了众多高质量数据集，推动了该领域的进展。然而，现有公开数据集主要由孤立的文本-视频（T-V）对组成，无法建模片段间关系。为解决这一局限，我们提出了 CI-VID，一个超越孤立 T2V 生成、面向文本-视频到视频（T&V2V）生成的数据集。CI-VID 包含超过 340,000 个样本，每个样本由语义连贯的视频序列及交错文本描述构成，这些描述同时捕捉片段级内容和片段间关系。为验证其有效性，我们设计了一个综合的多维度基准，涵盖人工评估、基于 VLM 的评估以及基于相似度的指标。实验结果表明，在 CI-VID 上训练的模型显著提升了多片段视频生成的准确性和内容一致性，从而能够创建具有平滑过渡和强语义连贯性的故事驱动内容。
英文原文摘要 1097 字符
Text-to-video (T2V) generation has recently attracted considerable attention, resulting in the development of numerous high-quality datasets that have propelled progress in this area. However, existing public datasets are primarily composed of isolated text-video (T-V) pairs and thus fail to model inter-clip relationships. To address this limitation, we introduce CI-VID, a dataset that moves beyond isolated T2V generation toward text-and-video-to-video (T&V2V) generation. CI-VID contains over 340,000 samples, each comprising a semantically coherent video sequence with interleaved text captions that capture both clip-level content and inter-clip relationships. To validate its effectiveness, we design a comprehensive, multi-dimensional benchmark incorporating human evaluation, VLM-based assessment, and similarity-based metrics. Experimental results demonstrate that models trained on CI-VID significantly improve both accuracy and content consistency in multi-clip video generation. This enables the creation of story-driven content with smooth transitions and strong semantic coherence.