teras.data_utils.create_gain_dataset

teras.data_utils.create_gain_dataset#

teras.data_utils.create_gain_dataset(x, seed=1337)[source]#

Creates a torch dataloader compatible with the GAIN architecture. The resultant dataset produces a tuple consisting of a batch of data for generator and another batch of data for the discriminator.

Parameters:
  • x – Dataset to use for training.

  • seed (int) – int, seed to use in shuffling. Defaults to 1337.