if nb >= batch_num:
break
labels = np.array(labels)
labels = labels.reshape(-1)
np.savetxt(os.path.join(p, 'label.txt'), labels)
def read_data_MNIST(p='./benchmark/Fashion_MNIST'):
labels = np.loadtxt(os.path.join(p, 'label.txt'))
imgs = []
for i in range(1, 161):
img_path = os.path.join(p, '{}.png'.format(i))
img = imread(img_path)
#print(img.shape)
img = img[:, :, 0]
imgs.append(img)
imgs = np.array(imgs)
print('X of shape {} --- Y of shape {}'.format(imgs.shape, labels.shape))
print('Data from {}'.format(p))
data = DATA(imgs, labels)
return data
if __name__ == '__main__':
#read_data_MNIST()
DL = DataLoader("Fashion_MNIST", 32, 28)
_, dl, _, _ = DL.Fashion_MNIST()
gen_benchmark(dl, p='./benchmark/Fashion_MNIST')
