ml.optimizers.gan

GAN model optimizer wrapper.

This wrapper allows downstream users to set different optimizers or optimizer parameters for the generator and discriminator of a GAN.

This class is used by the GAN trainer interface and shouldn’t be used elsewhere.

class ml.optimizers.gan.GenerativeAdversarialNetworkOptimizerConfig(name: str = '???', generator: Any = '???', discriminator: Any = '???')

Bases: BaseOptimizerConfig

generator: Any = '???'

discriminator: Any = '???'

classmethod update(config: DictConfig) → DictConfig

Runs post-construction config update.

Parameters:

config – The config to update

classmethod resolve(config: GenerativeAdversarialNetworkOptimizerConfig) → None

Runs post-construction config resolution.

Parameters:

config – The config to resolve

class ml.optimizers.gan.GenerativeAdversarialNetworkOptimizer(config: GenerativeAdversarialNetworkOptimizerConfig)

Bases: BaseOptimizer[GenerativeAdversarialNetworkOptimizerConfig, Optimizer]

get(model: Module) → Optimizer

Given a base module, returns an optimizer.

Parameters:

model – The model to get an optimizer for

Returns:

The constructed optimizer