import numpy
import theano
import theano.tensor as T
import os
import urllib
import gzip
import cPickle as pickle
def mnist_generator(data, batch_size):
images, targets = data
images = images.astype(theano.config.floatX)
targets = targets.astype('int32')
def get_epoch():
rng_state = numpy.random.get_state()
numpy.random.shuffle(images)
numpy.random.set_state(rng_state)
numpy.random.shuffle(targets)
image_batches = images.reshape(-1, batch_size, 784)
target_batches = targets.reshape(-1, batch_size)
for i in xrange(len(image_batches)):
yield (numpy.copy(image_batches[i]), numpy.copy(target_batches[i]))
return get_epoch
def load(batch_size, test_batch_size):
filepath = '/tmp/mnist.pkl.gz'
url = 'http://www.iro.umontreal.ca/~lisa/deep/data/mnist/mnist.pkl.gz'
if not os.path.isfile(filepath):
print "Couldn't find MNIST dataset in /tmp, downloading..."
urllib.urlretrieve(url, filepath)
with gzip.open('/tmp/mnist.pkl.gz', 'rb') as f:
train_data, dev_data, test_data = pickle.load(f)
return (
mnist_generator(train_data, batch_size),
mnist_generator(dev_data, test_batch_size),
mnist_generator(test_data, test_batch_size)
)
