specific method · filed under model architecture
Global attention
Attention spanning the global context, used alongside sliding-window attention in the described hybrid architecture.
Also called Global Attention (GA).
- sources
- 4
- models
- 2
- lab adopt it
- 1
- strongest
- core
How sources treat it
One count per evidence span, weakest treatment to strongest.
used 3core 1
Documented in
Evidence
4 spans quoted from the sources, strongest treatment first.
global attention
coremodel architecturein MiMo-V2-FlashXiaomi
Interleaves Sliding Window Attention (SWA) and Global Attention (GA) with a 6:1 ratio
usedmodel architecturein MiMo-V2.5-ProXiaomi
interleaving Sliding Window Attention (SWA) and Global Attention (GA) with a 5:1 ratio
usedmodel architecturein MiMo-V2.5Xiaomi
interleaving Sliding Window Attention (SWA) and Global Attention (GA) with a 5:1 ratio and 128 sliding window
usedmodel architecturein MiMo-V2.5Xiaomi
Filed alongside
Other methods under model architecture :: token mixer :: softmax attention :: global attention.