def build_finetune_functions(self, train_set, valid_set, test_set, batch_size, learning_rate):
'''
Generates a function `train` that implements one step of
finetuning, a function `validate` that computes the error on a
batch from the validation set, and a function `test` that
computes the error on a batch from the testing set
:type datasets: list of pairs of theano.tensor.TensorType
:param datasets: It is a list that contain all the datasets;
the has to contain three pairs, `train`,
`valid`, `test` in this order, where each pair
is formed of two Theano variables, one for the
datapoints, the other for the labels
:type batch_size: int
:param batch_size: size of a minibatch
:type learning_rate: float
:param learning_rate: learning rate used during finetune stage
'''
(train_set_x, train_set_y) = shared_dataset(train_set)
(valid_set_x, valid_set_y) = shared_dataset(valid_set)
(test_set_x, test_set_y) = shared_dataset(test_set)
# compute number of minibatchs for training, validataion and testing
n_valid_batchs = valid_set_x.get_value(borrow=True).shape[0]
n_valid_batchs /= batch_size
n_test_batchs = test_set_x.get_value(borrow=True).shape[0]
n_test_batchs /= batch_size
index = T.lscalar('index') # index to a [mini]batch
gparams = T.grad(self.finetune_cost, self.params)
updates = []
for param, gparams in zip(self.params, gparams):
updates.append((param, param - learning_rate * gparams))
train_fn = theano.function(inputs=[index], outputs=self.finetune_cost, updates=updates,
givens={self.x: train_set_x[index*batch_size:(index+1)*batch_size],
self.y: train_set_y[index*batch_size:(index+1)*batch_size]
}
)
test_score_i = theano.function([index], self.errors,
givens={self.x: valid_set_x[index*batch_size:(index+1)*batch_size],
self.y: valid_set_y[index*batch_size:(index+1)*batch_size]}
)
valid_score_i = theano.function([index], self.errors,
givens={self.x: test_set_x[index*batch_size:(index+1)*batch_size],
self.y: test_set_y[index*batch_size:(index+1)*batch_size]}
)
def valid_score():
return [valid_score_i(i) for i in xrange(n_valid_batchs)]
def test_score():
return [test_score_i(i) for i in xrange(n_test_batchs)]
return train_fn, valid_score, test_score
def test_DBN(finetune_lr=0.1, pretraining_epochs=100,
pretrain_lr=0.01, k=1, training_epochs=1000,
dataset='mnist.pkl.gz', batch_size=10):
"""
Demonstrates how to train and test a Deep Belief Network.
This is demonstrated on MNIST.
:type finetune_lr: float
:param finetune_lr: learning rate used in the finetune stage
:type pretraining_epochs: int
:param pretraining_epochs: number of epoch to do pretraining
:type pretrain_lr: float
:param pretrain_lr: learning rate to be used during pre-training
:type k: int
:param k: number of Gibbs steps in CD/PCD
:type training_epochs: int
:param training_epochs: maximal number of iterations ot run the optimizer
:type dataset: string
:param dataset: path the the pickled dataset
:type batch_size: int
:param batch_size: the size of a minibatch
"""
with gzip.open('/home/aurora/workspace/PycharmProjects/data/MNIST/mnist.pkl.gz', 'rb') as f:
train_set, validate_set, test_set = cPickle.load(f)
train_set_x, train_set_y = shared_dataset(train_set, borrow=True)
valid_set_x, valid_set_y = shared_dataset(validate_set, borrow=True)
test_set_x, test_set_y = shared_dataset(test_set, borrow=True)
# compute number of minibatches for training, validation and testing
n_train_batches = train_set_x.get_value(borrow=True).shape[0] / batch_size
# numpy random generator
numpy_rng = numpy.random.RandomState(123)
print '... building the model'
# construct the Deep Belief Network
dbn = DBN(numpy_rng=numpy_rng, n_ins=28 * 28,
hidden_layers_sizes=[1000, 1000, 1000],
n_outs=10)
# start-snippet-2
#########################
# PRETRAINING THE MODEL #
#########################
print '... getting the pretraining functions'
pretraining_fns = dbn.pretraining_functions(train_set_x=train_set_x,
batch_size=batch_size,
k=k)
print '... pre-training the model'
start_time = timeit.default_timer()
## Pre-train layer-wise
for i in xrange(dbn.n_layers):
# go through pretraining epochs
for epoch in xrange(pretraining_epochs):
# go through the training set
c = []
for batch_index in xrange(n_train_batches):
c.append(pretraining_fns[i](index=batch_index,
lr=pretrain_lr))
print 'Pre-training layer %i, epoch %d, cost ' % (i, epoch),
print numpy.mean(c)
end_time = timeit.default_timer()
# end-snippet-2
print >> sys.stderr, ('The pretraining code for file ' +
os.path.split(__file__)[1] +
' ran for %.2fm' % ((end_time - start_time) / 60.))
########################
# FINETUNING THE MODEL #
########################
# get the training, validation and testing function for the model
print '... getting the finetuning functions'
train_fn, validate_model, test_model = dbn.build_finetune_functions(
datasets=datasets,
batch_size=batch_size,
learning_rate=finetune_lr
)
print '... finetuning the model'
# early-stopping parameters
patience = 4 * n_train_batches # look as this many examples regardless
patience_increase = 2. # wait this much longer when a new best is
# found
improvement_threshold = 0.995 # a relative improvement of this much is
# considered significant
validation_frequency = min(n_train_batches, patience / 2)
# go through this many
# minibatches before checking the network
# on the validation set; in this case we
# check every epoch
best_validation_loss = numpy.inf
test_score = 0.
start_time = timeit.default_timer()
done_looping = False
epoch = 0
while (epoch < training_epochs) and (not done_looping):
epoch = epoch + 1
for minibatch_index in xrange(n_train_batches):
minibatch_avg_cost = train_fn(minibatch_index)
iter = (epoch - 1) * n_train_batches + minibatch_index
if (iter + 1) % validation_frequency == 0:
validation_losses = validate_model()
this_validation_loss = numpy.mean(validation_losses)
print(
'epoch %i, minibatch %i/%i, validation error %f %%'
% (
epoch,
minibatch_index + 1,
n_train_batches,
this_validation_loss * 100.
)
)
# if we got the best validation score until now
if this_validation_loss < best_validation_loss:
#improve patience if loss improvement is good enough
if (
this_validation_loss < best_validation_loss *
improvement_threshold
):
patience = max(patience, iter * patience_increase)
# save best validation score and iteration number
best_validation_loss = this_validation_loss
best_iter = iter
# test it on the test set
test_losses = test_model()
test_score = numpy.mean(test_losses)
print((' epoch %i, minibatch %i/%i, test error of '
'best model %f %%') %
(epoch, minibatch_index + 1, n_train_batches,
test_score * 100.))
if patience <= iter:
done_looping = True
break
end_time = timeit.default_timer()
print(
(
'Optimization complete with best validation score of %f %%, '
'obtained at iteration %i, '
'with test performance %f %%'
) % (best_validation_loss * 100., best_iter + 1, test_score * 100.)
)
print >> sys.stderr, ('The fine tuning code for file ' +
os.path.split(__file__)[1] +
' ran for %.2fm' % ((end_time - start_time)
/ 60.))
