CIFAR-100 WideResNet 40-4

class deepobs.cifar100.cifar100_wrn404.set_up[source]

Function providing the functionality for the WideResNet 40-4 architecture on CIFAR-100.

This function is a wrapper for the cifar100_wrn.set_up to create the 40-4 variant of the WideResNet.

The training setting in the paper were: Batch size of 128, weight decay of 0.0005, total training time of 200 epochs, with a learning rate schedule of 0.1, 0.02 after 60 epochs, 0.004 after 120 epochs and 0.0008 after 160 epochs. Training was done using Nesterov momentum with a momentum parameter of 0.9.

Parameters:
  • batch_size (int) -- Batch size of the data points.
  • weight_decay (float) -- Weight decay factor. Defaults to 0.0005.
Returns:

Setup class for WideResNets on CIFAR-100, cifar100_wrn.set_up.
Return type:

cifar100_wrn.set_up