implementation detail · filed under model architecture
Exact coordinate minimization
An implementation step for Quantile Balancing that solves each subproblem exactly in closed form.
- 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.
each subproblem admits a closed-form exact solution.
usedoptimizationin Kimi K3Moonshot AI
Filed alongside
Other methods under model architecture :: channel mixer :: mixture of experts :: expert load balancing.
Auxiliary-loss-free load balancingQuantile BalancingHistogram-based quantile estimationAuxiliary-loss load balancingExpert bias update factorExpert Parallelism Load BalancingExponential moving average of estimated quantilesformHCLoad balancingMaxVioModality-specific auxiliary-loss-free load balancingPersistent load balancingPooled-global-batch quantile estimationRound-robin load balancing