'valid_set': BatchIterator('continuous'),
'test_set': BatchIterator('continuous')
},
device="/cpu:0")
# Create Network
#---------------
# we use a batch_size of 64 and use the dataset.datum shape to
# obtain the shape of 1 observation and create the input shape
K = 8
my_layer = layers.custom_layer(ops.Dense, ops.BatchNorm, ops.Activation)
dnn = sknet.network.Network(name='simple_model')
dnn.append(ops.RandomCrop(dataset.images, (28, 28)))
dnn.append(
layers.Conv2D(dnn[-1], [(64, 3, 3), {
'b': None,
'pad': 'same'
}], [[0, 2, 3]], [0.01]))
dnn.append(my_layer(dnn[-1], [10, {'b': None}], [0], [tf.identity]))
dnn.append(
layers.Conv2DPool(dnn[-2], [(192, 3, 3), {
'b': None,
'pad': 'same'
}], [[0, 2, 3]], [0.01], [(1, 2, 2)]))
dnn.append(my_layer(dnn[-1], [10, {'b': None}], [0], [tf.identity]))
return tf.abs(tensor[:,c])/norm
distances = tf.map_fn(doit,tf.range(tensor.shape.as_list()[1]),
dtype=tf.float32)
return tf.reduce_min(distances,0)
# Create Network
#---------------
dnn = sknet.Network(name='simple_model')
if DATA_AUGMENTATION:
dnn.append(ops.RandomAxisReverse(dataset.images,axis=[-1]))
dnn.append(ops.RandomCrop(dnn[-1],(28,28),seed=10))
start_op = 2
else:
dnn.append(dataset.images)
start_op = 1
if MODEL=='cnn':
sknet.networks.ConvSmall(dnn,dataset.n_classes)
elif MODEL=='dense':
dnn.append(sknet.ops.Dense(dnn[-1],4096))
dnn.append(sknet.ops.BatchNorm(dnn[-1],0))
dnn.append(sknet.ops.Activation(dnn[-1],0.1))
dnn.append(sknet.ops.Dense(dnn[-1],2048))
dnn.append(sknet.ops.BatchNorm(dnn[-1],0))
dnn.append(sknet.ops.Activation(dnn[-1],0.1))
'train_set': 'random_see_all',
'test_set': 'continuous'
})
dataset.create_placeholders(iterator, device="/cpu:0")
# Create Network
#---------------
# we use a batch_size of 64 and use the dataset.datum shape to
# obtain the shape of 1 observation and create the input shape
dnn = sknet.Network(name='simple_model')
dnn.append(ops.RandomAxisReverse(dataset.images, axis=[-1]))
dnn.append(ops.RandomCrop(dnn[-1], (28, 28)))
if MODEL == 'smallcnn':
sknet.networks.ConvSmall(dnn, dataset.n_classes)
all_layers = dnn[2:-1].as_list()
elif MODEL == 'largecnn':
sknet.networks.ConvLarge(dnn, dataset.n_classes)
all_layers = dnn[2:-1].as_list()
else:
sknet.networks.Resnet(dnn, dataset.n_classes, D=2, W=2)
all_layers = [
layer for layer in dnn[2:-1] if type(layer) == sknet.ops.Merge
]
prediction = dnn[-1]