Hands-On Generative Adversarial Networks with Keras

Initial setup

The initial setup for training can be done using the following steps:

1. Add a summary writer for logging using the following code:

    # add a summary writer for logging     logger = SummaryWriter(out_dir)    logger.add_scalar('d_lr', lr_d, 0)    logger.add_scalar('g_lr', lr_g, 0)

2. Load training and validation data and instantiate data iterators:

    # load training and validation data and instantiate data iterators    train_data = get_data(data_filepath, 'train')    val_data = get_data(data_filepath, 'valid')    data_iterator = iterate_minibatches(train_data, batch_size)    val_data_iterator = iterate_minibatches(val_data, n_checkpoint_samples)

3. Load the sample validation data to use as a reference during training:

 # sample validation data to use ...