implementation detail · filed under model architecture
Micro-batch-level shared-state lifetime management
A shared-state management detail that retains states until their final consumer completes, then releases them to limit memory overhead.
Also called Micro-batch-level shared-state management.
- 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.
Shared states are retained until their final consumer has completed and are then released promptly to limit additional memory overhead
usedunclearin DeepSeek-V4.1-FlashDeepSeek
Filed alongside
Other methods under model architecture :: token mixer.