Don't Decay the Learning Rate, Increase the Batch Size.

Samuel L. Smith Pieter-Jan Kindermans Chris Ying Quoc V. Le

Published in: ICLR (Poster) (2018)

Keyphrases

learning rate
batch size
convergence rate

learning algorithm
convergence speed
batch mode

batch processing
poisson process
single item

multi objective
multi class
simulated annealing