|
@@ -1,19 +1,24 @@
|
|
|
from mnist import MNIST
|
|
|
import numpy as np
|
|
|
|
|
|
-def load_training_samples():
|
|
|
+def load_samples(training):
|
|
|
"""
|
|
|
+ training : load training data if true, load testing data otherwise
|
|
|
+
|
|
|
Return np_images, np_expected
|
|
|
where
|
|
|
np_impages is a np.array of 784 x 60 000
|
|
|
np_expected is a np.array of 10 x 60 000
|
|
|
"""
|
|
|
- mndata = MNIST('../../resources/download')
|
|
|
+ mndata = MNIST('../resources/download')
|
|
|
|
|
|
images = [[]] # Contains vectors of 784 pixels image
|
|
|
labels = [] # Contains expected response for each image
|
|
|
|
|
|
- images, labels = mndata.load_training()
|
|
|
+ if (training):
|
|
|
+ images, labels = mndata.load_training()
|
|
|
+ else:
|
|
|
+ images, labels = mndata.load_testing()
|
|
|
|
|
|
np_images = np.array(images, dtype=np.float64)
|
|
|
|
|
@@ -26,3 +31,9 @@ def load_training_samples():
|
|
|
np_expected[k][label] = 1.0
|
|
|
|
|
|
return np.transpose(np_images), np.transpose(np_expected)
|
|
|
+
|
|
|
+def load_training_samples():
|
|
|
+ return load_samples(training = True)
|
|
|
+
|
|
|
+def load_testing_samples():
|
|
|
+ return load_samples(training = False)
|