Pretrain, finetune ANY AI model of ANY size on 1 or 10,000+ GPUs with zero code changes.
Fix `StochasticWeightAveraging` with infinite epochs (#21396)
* implement special case max_epoch==-1 * add testing * changelog --------- Co-authored-by: Bhimraj Yadav <[email protected]>
N
Nicki Skafte Detlefsen committed
f3f6605e1ad91c8e1a5d969eba678b7a55af3d78
Parent: 3876cc5
Committed by GitHub <[email protected]>
on 12/15/2025, 7:03:25 PM