TS-Attn dynamically separates and rearranges attention in existing text-to-video models to improve temporal consistency and prompt adherence for videos with multiple sequential actions.
As shown in Figure 9, the attention distributions of different actions in TS-Attn are clearly separated along the temporal sequence
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
TS-Attn: Temporal-wise Separable Attention for Multi-Event Video Generation
TS-Attn dynamically separates and rearranges attention in existing text-to-video models to improve temporal consistency and prompt adherence for videos with multiple sequential actions.