Research Article

AdaCN: An Adaptive Cubic Newton Method for Nonconvex Stochastic Optimization

Table 1

Summaries of the settings used in experiments.

TaskDatasetNetwork typeArchitectureOptimizer

Image classificationMnistConvolutional neural networkLeNetAdaCN, Apollo, Adam, SGD, AdaBound
CIFAR10
CIFAR100
VGG11, ResNet34, DenseNet121
Language modelingPenn TreebankRecurrent1, 2, 3-layer LSTM