implementation detail · filed under optimization
Exponential moving average of checkpoints
Averages recent checkpoints, here over the 10 most recent, during progressive long-context extension training.
Also called exponential moving average (EMA).
- 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.
we apply an exponential moving average (EMA) over the 10 most recent checkpoints
usedoptimizationin Laguna XS.2Poolside
Filed alongside
Other methods under optimization :: training stability.
Weight paddingActivation or logit clippingAnti-hallucination trainingBehavioral regularizationCross-replica model-weight hash consistency checksElevated constant-learning-rate training stability stress testGradient clippingLearning-rate elevation for training stability stress testingLoss masking for excessively stale tokensOff-policy sample filteringPer-token regularization for off-policy RLSafeguards against training drift and reward hackingSoft droppingTraining stability stress testingWeight clippingWeight decay coupled to learning-rate squared