def residual_network(n):
'''
n: the network contains 6 * n + 2 layers, including 6 * n + 1 convolution layers and 1 affine layer
please refer to the paper for details
'''
def normalized_convolution(kernel_shape, kernel_number, stride, pad, activate=None):
module = builder.Sequential(
builder.Convolution(kernel_shape, kernel_number, stride, pad),
builder.SpatialBatchNormalization()
)
if activate:
module.append(getattr(builder, activate)())
return module
def residual(kernel_number, project=False):
if project:
module = builder.Add(
builder.Sequential(
normalized_convolution((3, 3), kernel_number, (2, 2), (1, 1), 'ReLU'),
normalized_convolution((3, 3), kernel_number, (1, 1), (1, 1))
),
builder.Sequential(
builder.Pooling('avg', (2, 2), (2, 2)),
builder.Convolution((1, 1), kernel_number)
)
)
else:
module = builder.Add(
builder.Sequential(
normalized_convolution((3, 3), kernel_number, (1, 1), (1, 1), 'ReLU'),
normalized_convolution((3, 3), kernel_number, (1, 1), (1, 1))
),
builder.Identity()
)
return module
network = builder.Sequential(
builder.Reshape((3, 32, 32)),
normalized_convolution((3, 3), 16, (1, 1), (1, 1), 'ReLU')
)
for i in range(n):
network.append(residual(16))
network.append(residual(32, project=True))
for i in range(n-1):
network.append(residual(32))
network.append(residual(64, project=True))
for i in range(n-1):
network.append(residual(64))
network.append(builder.Pooling('avg', (8, 8)))
network.append(builder.Reshape((64,)))
network.append(builder.Affine(10))
return network
def residual(kernel_number, project=False):
if project:
module = builder.Add(
builder.Sequential(
normalized_convolution((3, 3), kernel_number, (2, 2),
(1, 1), 'ReLU'),
normalized_convolution((3, 3), kernel_number, (1, 1),
(1, 1))),
builder.Sequential(builder.Pooling('avg', (2, 2), (2, 2)),
builder.Convolution((1, 1), kernel_number)))
else:
module = builder.Add(
builder.Sequential(
normalized_convolution((3, 3), kernel_number, (1, 1),
(1, 1), 'ReLU'),
normalized_convolution((3, 3), kernel_number, (1, 1),
(1, 1))), builder.Identity())
return module
def main(args):
# Define a convolutional neural network the same as above
net = builder.Sequential(
builder.Convolution((7, 7), 32),
builder.ReLU(),
builder.Pooling('max', (2, 2), (2, 2)),
builder.Reshape((flattened_input_size,))
builder.Affine(hidden_size),
builder.Affine(num_classes),
)
# Cast the definition to a model compatible with minpy solver
model = builder.Model(net, 'softmax', (3 * 32 * 32,))
data = get_CIFAR10_data(args.data_dir)
train_dataiter = NDArrayIter(data['X_train'],
data['y_train'],
batch_size=batch_size,
shuffle=True)
test_dataiter = NDArrayIter(data['X_test'],
data['y_test'],
batch_size=batch_size,
shuffle=False)
solver = Solver(model,
train_dataiter,
test_dataiter,
num_epochs=10,
init_rule='gaussian',
init_config={
'stdvar': 0.001
},
update_rule='sgd_momentum',
optim_config={
'learning_rate': 1e-3,
'momentum': 0.9
},
verbose=True,
print_every=20)
solver.init()
solver.train()
'''
Network In Network
Reference:
Min Lin, Qiang Chen, Shuicheng Yan, Network In Network
'''
network_in_network = builder.Sequential(
builder.Reshape((3, 32, 32)),
builder.Convolution((5, 5), 192, pad=(2, 2)),
builder.ReLU(),
builder.Convolution((1, 1), 160),
builder.ReLU(),
builder.Convolution((1, 1), 96),
builder.ReLU(),
builder.Pooling('max', (3, 3), (2, 2), (1, 1)),
builder.Dropout(0.5),
builder.Convolution((5, 5), 192, pad=(2, 2)),
builder.ReLU(),
builder.Convolution((1, 1), 192),
builder.ReLU(),
builder.Convolution((1, 1), 192),
builder.ReLU(),
builder.Pooling('avg', (3, 3), (2, 2), (1, 1)),
builder.Dropout(0.5),
builder.Convolution((3, 3), 192, pad=(1, 1)),
builder.ReLU(),
builder.Convolution((1, 1), 192),
builder.ReLU(),
builder.Convolution((1, 1), 10),
builder.ReLU(),
import minpy.nn.model_builder as builder
'''
Network In Network
Reference:
Min Lin, Qiang Chen, Shuicheng Yan, Network In Network
'''
network_in_network = builder.Sequential(
builder.Reshape((3, 32, 32)), builder.Convolution((5, 5), 192, pad=(2, 2)),
builder.ReLU(), builder.Convolution((1, 1), 160), builder.ReLU(),
builder.Convolution((1, 1), 96), builder.ReLU(),
builder.Pooling('max', (3, 3), (2, 2)), builder.Dropout(0.5),
builder.Convolution((5, 5), 192, pad=(2, 2)), builder.ReLU(),
builder.Convolution((1, 1), 192), builder.ReLU(),
builder.Convolution((1, 1), 192), builder.ReLU(),
builder.Pooling('avg', (3, 3), (2, 2)), builder.Dropout(0.5),
builder.Convolution((3, 3), 192, pad=(1, 1)), builder.ReLU(),
builder.Convolution((1, 1), 192), builder.ReLU(),
builder.Convolution((1, 1), 10), builder.ReLU(),
builder.Pooling('avg', (8, 8)), builder.Reshape((10, )))
'''
Residual Network
Reference:
Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun. "Deep Residual Learning for Image Recognition"
'''
def residual_network(n):
'''
n: the network contains 6 * n + 2 layers, including 6 * n + 1 convolution layers and 1 affine layer