Research map / Video / Video generation
TS-Attn: Temporal-wise Separable Attention for Multi-Event Video Generation
Hongyu Zhang, Yufan Deng, Zilin Pan, Peng-Tao Jiang, Bo Li, Qibin Hou, Zhen Dong, Zhiyang Dou, Daquan Zhou
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add TS-Attn | Wan2.2-A14B · StoryEval-Bench T2V (GPT-4o) | StoryEval score (Average): 48.3% -> 56.2% | Improved |
| add TS-Attn | Wan2.1-14B · StoryEval-Bench T2V (GPT-4o) | StoryEval score (Average): 37.6% -> 50.2% | Improved |
| add TS-Attn | CogVideoX-5B · StoryEval-Bench T2V (GPT-4o) | StoryEval score (Average): 16.4% -> 25.8% | Improved |
| add TS-Attn | Wan2.1-1.3B · StoryEval-Bench T2V (GPT-4o) | StoryEval score (Average): 29.1% -> 37.6% | Improved |
| add TS-Attn | Wan2.2-I2V-A14B · StoryEval-Bench-I2V (GPT-4o) | StoryEval score (Average): 47.5% -> 54.4% | Improved |
| add TS-Attn | CogVideoX-I2V-5B · StoryEval-Bench-I2V (GPT-4o) | StoryEval score (Average): 19.6% -> 28.3% | Improved |
| add TS-Attn instead of multi-prompt MEVG | Wan2.2-A14B · StoryEval-Bench T2V (GPT-4o) | StoryEval score (Average): 43.1% -> 56.2% | Improved |
| add TS-Attn instead of multi-prompt DiTCtrl | Wan2.2-A14B · StoryEval-Bench T2V (GPT-4o) | StoryEval score (Average): 46.5% -> 56.2% | Improved |
| add TS-Attn (inference cost) | Wan2.2-A14B · StoryEval-Bench inference time (A100) | latency (s): 846 -> 863 | No clear effect |
| add event-aware attention modulation (EAM) (ablation) | Wan2.1-14B · StoryEval-Bench T2V (GPT-4o) | StoryEval score (Average): 37.6% -> 46.4% | Improved |
| add event-aware attention modulation (EAM) (ablation) | CogVideoX-5B · StoryEval-Bench T2V (GPT-4o) | StoryEval score (Average): 16.4% -> 22.9% | Improved |
| add event-aware attention modulation (EAM) (ablation) | Wan2.2-A14B · StoryEval-Bench T2V (GPT-4o) | StoryEval score (Average): 48.3% -> 51.9% | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- CogVideoX-5B — CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer (ICLR 2025) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Mind the Time: Temporally-Controlled Multi-Event Video Generation (CVPR 2025 · similarity 0.88)
- DiTCtrl: Exploring Attention Control in Multi-Modal Diffusion Transformer for Tuning-Free Multi-Prompt Longer Video Generation (CVPR 2025 · similarity 0.87)
- VSTAR: Generative Temporal Nursing for Longer Dynamic Video Synthesis (ICLR 2025 · similarity 0.87)
- ByTheWay: Boost Your Text-to-Video Generation Model to Higher Quality in a Training-free Way (CVPR 2025 · similarity 0.86)
- Enhancing Temporal Understanding in Video-LLMs through Stacked Temporal Attention in Vision Encoders (NeurIPS 2025 · similarity 0.86)
- EventLens: Event-Structure Reinforcement Learning for Video Understanding (NeurIPS 2026 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
