def base(train_dp, valid_dp, logger, learning_rate):
# learning_rate = 0.01
rng = numpy.random.RandomState([2016,02,26])
max_epochs = 1000
cost = CECost()
stats = list()
test_dp = deepcopy(valid_dp)
train_dp.reset()
valid_dp.reset()
test_dp.reset()
# NETWORK TOPOLOGY:
model = MLP(cost=cost)
model.add_layer(Relu(idim=125, odim=125, irange=1.6, rng=rng))
model.add_layer(Softmax(idim=125, odim=19, rng=rng))
# define the optimiser, here stochasitc gradient descent
# with fixed learning rate and max_epochs
lr_scheduler = LearningRateFixed(
learning_rate=learning_rate, max_epochs=max_epochs)
optimiser = SGDOptimiser(lr_scheduler=lr_scheduler)
logger.info('Training started...')
tr_stats_b, valid_stats_b = optimiser.train(model, train_dp, valid_dp)
logger.info('Testing the model on test set:')
tst_cost, tst_accuracy = optimiser.validate(model, test_dp)
logger.info('ACL test set accuracy is %.2f %%, cost (%s) is %.3f' %
(tst_accuracy*100., cost.get_name(), tst_cost))
def get_cur_encoder_model(model, cur_layer_id):
cur_layer = model.layers[cur_layer_id]
assert isinstance(cur_layer, Linear), (
"Expected current layer to be Linear or its subclass")
if cur_layer_id == 0:
get_inputs = lambda x: (masking_noise(x), x)
else:
prev_layers = model.layers[:cur_layer_id]
prev_mds = MLP_fast(MSECost())
prev_mds.set_layers(prev_layers)
def get_inputs_noisy(x):
pure = prev_mds.fprop(x)
return masking_noise(pure), pure
get_inputs = get_inputs_noisy
if cur_layer_id == last_layer:
assert isinstance(cur_layer, Softmax), (
"final layer must be softmax for MNIST digits classification"
)
#here it automatically matches output of previous layer
get_targets = lambda inputs, t, pure: t
cur_model = MLP_fast(CECost())
cur_model.add_layer(cur_layer)
else:
get_targets = lambda inputs, t, pure: pure
cur_model = MLP_fast(MSECost())
cur_model.add_layer(cur_layer)
#echo the output of the current layer
cur_model.add_layer(Linear(cur_layer.odim, cur_layer.idim))
return cur_model, get_inputs, get_targets
def get_cur_discr_model(model, cur_layer_id):
cur_layer = model.layers[cur_layer_id]
assert isinstance(cur_layer, Linear), (
"Expected current layer to be Linear or its subclass")
get_targets = lambda inputs, t, pure: t
if cur_layer_id == 0:
get_inputs = lambda x: (x, None)
else:
prev_layers = model.layers[:cur_layer_id]
prev_mds = MLP_fast(MSECost())
prev_mds.set_layers(prev_layers)
get_inputs = lambda x: (prev_mds.fprop(x), None)
last_layer = cur_layer_id == len(model.layers) - 1
cur_model = MLP_fast(CECost())
cur_model.add_layer(cur_layer)
if last_layer:
assert isinstance(cur_layer, Softmax), (
"final layer must be softmax for MNIST digits classification"
)
#here it automatically matches output of previous layer
else:
#get final layer for the MNIST dataset
cur_model.add_layer(Softmax(cur_layer.odim, 10))
return cur_model, get_inputs, get_targets
from mlp.layers import MLP, Linear, Sigmoid, Softmax #import required layer types from mlp.layers import * from mlp.optimisers import SGDOptimiser #import the optimiser from mlp.costs import CECost #import the cost we want to use for optimisation from mlp.schedulers import LearningRateFixed logger = logging.getLogger() logger.setLevel(logging.INFO) rng = numpy.random.RandomState([2015,10,10]) #some hyper-parameters nhid = 100 learning_rate = 0.07 max_epochs = 30 cost = CECost() stats = list() test_dp = deepcopy(valid_dp) train_dp.reset() valid_dp.reset() test_dp.reset() #define the model model = MLP(cost=cost) #model.add_layer(ComplexLinear(idim=125, odim=125, irange=1.6, rng=rng)) #model.add_layer(Sigmoid(idim=2*125, odim=125, irange=1.6, rng=rng)) model.add_layer(Sigmoid(idim=125, odim=125, irange=1.6, rng=rng)) model.add_layer(Softmax(idim=125, odim=19, rng=rng))
logger = logging.getLogger()
logger.setLevel(logging.INFO)
from mlp.layers import MLP, Sigmoid, Linear, Softmax, Relu #import required layer types
from mlp.conv import ConvLinear, ConvRelu, ConvSigmoid
from mlp.maxpooling import ConvMaxPool2D
from mlp.optimisers import SGDOptimiser, Optimiser #import the optimiser
from mlp.dataset import MNISTDataProvider #import data provider #Ruslan Burakov - s1569105
from mlp.costs import CECost, MSECost #import the cost we want to use for optimisation
from mlp.schedulers import LearningRateFixed
rng = numpy.random.RandomState([2015, 10, 10])
# define the model structure, here just one linear layer
# and mean square error cost
cost = CECost()
tsk8_2_model = MLP(cost=cost)
"""
num_inp_feat_maps,
num_out_feat_maps,
image_shape=(28, 28),
kernel_shape=(5, 5),
stride=(1, 1),
irange=0.2,
rng=None,
conv_fwd=my_conv_fwd,
conv_bck=my_conv_bck,
conv_grad=my_conv_grad)
"""
tsk8_2_model.add_layer(
ConvRelu(num_inp_feat_maps=1,
from mlp.layers import MLP, Linear, Sigmoid, Softmax #import required layer types from mlp.layers import * from mlp.optimisers import SGDOptimiser #import the optimiser from mlp.costs import CECost #import the cost we want to use for optimisation from mlp.schedulers import LearningRateFixed logger = logging.getLogger() logger.setLevel(logging.INFO) rng = numpy.random.RandomState([2015, 10, 10]) #some hyper-parameters nhid = 100 learning_rate = 0.07 max_epochs = 30 cost = CECost() stats = list() test_dp = deepcopy(valid_dp) train_dp.reset() valid_dp.reset() test_dp.reset() #define the model model = MLP(cost=cost) #model.add_layer(ComplexLinear(idim=125, odim=125, irange=1.6, rng=rng)) #model.add_layer(Sigmoid(idim=2*125, odim=125, irange=1.6, rng=rng)) model.add_layer(Sigmoid(idim=125, odim=125, irange=1.6, rng=rng)) model.add_layer(Softmax(idim=125, odim=19, rng=rng))
logger = logging.getLogger()
logger.setLevel(logging.INFO)
from mlp.layers import MLP, Sigmoid, Linear, Softmax #import required layer types
from mlp.conv import ConvLinear, ConvRelu, ConvSigmoid
from mlp.maxpooling import ConvMaxPool2D
from mlp.optimisers import SGDOptimiser, Optimiser #import the optimiser
from mlp.dataset import MNISTDataProvider #import data provider #Ruslan Burakov - s1569105
from mlp.costs import CECost, MSECost #import the cost we want to use for optimisation
from mlp.schedulers import LearningRateFixed
rng = numpy.random.RandomState([2015, 10, 10])
# define the model structure, here just one linear layer
# and mean square error cost
tsk8_1_cost = CECost()
tsk8_1_model = MLP(cost=tsk8_1_cost)
"""
num_inp_feat_maps,
num_out_feat_maps,
image_shape=(28, 28),
kernel_shape=(5, 5),
stride=(1, 1),
irange=0.2,
rng=None,
conv_fwd=my_conv_fwd,
conv_bck=my_conv_bck,
conv_grad=my_conv_grad)
"""
tsk8_1_model.add_layer(
ConvSigmoid(num_inp_feat_maps=1,
