Experiment - Annotated Deep Learning Paper Implementations

DCGAN

This is an experiment training DCGAN model.

Install the labml-nn package

python

!pip install labml-nn

Imports

python


from labml import experiment
from labml_nn.gan.wasserstein.experiment import Configs

Create an experiment

python

experiment.create(name="mnist_wgan")

Initialize configurations

python

conf = Configs()

Set experiment configurations and assign a configurations dictionary to override configurations

python

experiment.configs(conf,
                   {
                       'discriminator': 'cnn',
                       'generator': 'cnn',
                       'label_smoothing': 0.01,
                       'generator_loss': 'wasserstein',
                       'discriminator_loss': 'wasserstein',
                   })

Start the experiment and run the training loop.

python

with experiment.start():
    conf.run()