arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2401.03048cs.CV
收藏

Latte:用于视频生成的潜在扩散Transformer

Latte: Latent Diffusion Transformer for Video Generation

Department of Data Science & AI, Faculty of Information Technology, Monash University(数据科学与人工智能系,信息科技学院,莫纳什大学);Shanghai AI Laboratory(上海人工智能实验室);Nanjing University of Posts and Telecommunications(南京邮电大学);S-Lab, Nanyang Technological University(南洋理工大学S实验室)

Xin Ma, Yaohui Wang, Xinyuan Chen, Gengyun Jia, Ziwei Liu, Yuan-Fang Li, Cunjian Chen, Yu Qiao

AI总结:

提出Latte,一种用于视频生成的新型潜在扩散Transformer,通过时空token提取与四种变体高效建模视频分布,在多个标准数据集上达到SOTA,并成功扩展至T2V任务。

摘要:

我们提出了Latte,一种用于视频生成的新型潜在扩散Transformer。Latte首先从输入视频中提取时空token,然后采用一系列Transformer块在潜在空间中对视频分布进行建模。为了对从视频中提取的大量token进行建模,从分解输入视频的空间和时间维度的角度,引入了四种高效变体。为了提高生成视频的质量,我们通过严格的实验分析确定了Latte的最佳实践,包括视频片段patch embedding、模型变体、timestep-class信息注入、时间位置embedding和学习策略。我们的综合评估表明,Latte在四个标准视频生成数据集(即FaceForensics、SkyTimelapse、UCF101和Taichi-HD)上取得了最先进的性能。此外,我们将Latte扩展到文本到视频生成(T2V)任务,Latte取得了与近期T2V模型具有竞争力的结果。我们坚信,Latte为未来将Transformer整合到视频生成扩散模型中的研究提供了有价值的见解。

英文摘要:

We propose Latte, a novel Latent Diffusion Transformer for video generation. Latte first extracts spatio-temporal tokens from input videos and then adopts a series of Transformer blocks to model video distribution in the latent space. In order to model a substantial number of tokens extracted from videos, four efficient variants are introduced from the perspective of decomposing the spatial and temporal dimensions of input videos. To improve the quality of generated videos, we determine the best practices of Latte through rigorous experimental analysis, including video clip patch embedding, model variants, timestep-class information injection, temporal positional embedding, and learning strategies. Our comprehensive evaluation demonstrates that Latte achieves state-of-the-art performance across four standard video generation datasets, i.e., FaceForensics, SkyTimelapse, UCF101, and Taichi-HD. In addition, we extend Latte to the text-to-video generation (T2V) task, where Latte achieves results that are competitive with recent T2V models. We strongly believe that Latte provides valuable insights for future research on incorporating Transformers into diffusion models for video generation.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑