def _(): base_model_fn = _fc_layer_norm_loss_fn([128, 128, 128, 10], tf.tanh) return base.DatasetModelTask( base_model_fn, datasets.get_image_datasets("cifar10", batch_size=128))
def _(): base_model_fn = _fc_batch_norm_loss_fn([64, 64, 64, 64, 64, 10], tf.nn.relu) return base.DatasetModelTask( base_model_fn, datasets.get_image_datasets("cifar10", batch_size=128))
def _(): base_model_fn = conv_ae_loss_fn([32, 32], [32, 32], 32, tf.nn.relu) dataset = datasets.get_image_datasets("mnist", batch_size=128) return base.DatasetModelTask(base_model_fn, dataset)
def _(): base_model_fn = _fc_dropout_loss_fn([128, 128, 10], tf.nn.relu, keep_probs=0.2) return base.DatasetModelTask( base_model_fn, datasets.get_image_datasets("cifar10", batch_size=128))
def _(): base_model_fn = fc_ae_loss_fn([128, 32, 128], tf.nn.relu) dataset = datasets.get_image_datasets("cifar10", batch_size=128) return base.DatasetModelTask(base_model_fn, dataset)
def _(): base_model_fn = conv_ae_loss_fn([32, 64], [64, 32], 8, tf.nn.relu) dataset = datasets.get_image_datasets("cifar10", batch_size=128) return base.DatasetModelTask(base_model_fn, dataset)
def _(): base_model_fn = ce_flatten_loss([32, 64, 64], tf.nn.relu, []) dataset = datasets.get_image_datasets( "food101_64x64", batch_size=64, shuffle_buffer=5000) return base.DatasetModelTask(base_model_fn, dataset)
def _(): base_model_fn = get_loss_fn(3, (1024, 1024)) dataset = datasets.get_image_datasets("cifar10", batch_size=64) return base.DatasetModelTask(base_model_fn, dataset)
def _(): base_model_fn = ce_flatten_loss([32, 16, 64], tf.nn.tanh, [32]) dataset = datasets.get_image_datasets("mnist", batch_size=32) return base.DatasetModelTask(base_model_fn, dataset)
def _(): base_model_fn = ce_flatten_loss([32, 64, 64], tf.nn.relu, []) dataset = datasets.get_image_datasets("cifar100", batch_size=128) return base.DatasetModelTask(base_model_fn, dataset)
def _(): # pylint: disable=missing-docstring base_model_fn = ce_pool_loss([32, 32, 32, 64, 64], tf.nn.relu, use_batch_norm=True) dataset = datasets.get_image_datasets("cifar10", batch_size=128) return base.DatasetModelTask(base_model_fn, dataset)
def _(): base_model_fn = get_loss_fn(9, layers=(128, 128)) dataset = datasets.get_image_datasets("mnist", batch_size=64) return base.DatasetModelTask(base_model_fn, dataset)
def _(): base_model_fn = fc_vae_loss_fn([128, 64], [64, 128], 32, tf.nn.relu) dataset = datasets.get_image_datasets( "food101_32x32", batch_size=256, shuffle_buffer=5000) return base.DatasetModelTask(base_model_fn, dataset)
def _(): base_model_fn = conv_vae_loss_fn([64, 128], [128, 64], 128, tf.nn.relu) dataset = datasets.get_image_datasets("mnist", batch_size=128) return base.DatasetModelTask(base_model_fn, dataset)
def _(): base_model_fn = ce_pool_loss([32, 64, 64], tf.nn.tanh) dataset = datasets.get_image_datasets("cifar10", batch_size=64) return base.DatasetModelTask(base_model_fn, dataset)
def _(): base_model_fn = three_layer_conv_vae_loss_fn([32, 64, 128], [128, 64, 32], 64, tf.nn.relu) dataset = datasets.get_image_datasets("cifar10", batch_size=128) return base.DatasetModelTask(base_model_fn, dataset)
def _(): base_model_fn = get_loss_fn(2, (2048, 2048)) dataset = datasets.get_image_datasets("mnist", batch_size=64) return base.DatasetModelTask(base_model_fn, dataset)