implementation detail · filed under model architecture
Factorized spatial-temporal video attention with temporal pooling
Video attention that separates spatial within-frame and temporal between-frame passes, with pooling to compress tokens over time.
- source
- 1
- model
- 1
- lab adopt it
- 1
- strongest
- used
How sources treat it
One count per evidence span, weakest treatment to strongest.
used 1
Documented in
Evidence
1 span quoted from the sources, strongest treatment first.
attention is factorized into intra-frame spatial and inter-frame temporal passes, and temporal pooling further compresses tokens along the time dimension.
usedmodel architecturein Kimi K3Moonshot AI
Filed alongside
Other methods under model architecture :: token mixer.