From: Martin Pitt <martin@piware.de>
Date: Sat, 29 Aug 2020 05:45:10 +0000 (+0200)
Subject: Rearange image vector
X-Git-Url: https://piware.de/gitweb/?p=handwriting-recognition.git;a=commitdiff_plain;h=8af4223121b60d5d67b7121d87c5c6fed01b58e7

Rearange image vector

Put each image into a column instead of a row, which works much better
with the standard formulation of backpropagation algorithms.
---

diff --git a/mnist.py b/mnist.py
index 1a3870f..4c2832c 100644
--- a/mnist.py
+++ b/mnist.py
@@ -11,8 +11,8 @@ def load(images_file, labels_file):
         rows = struct.unpack('>I', f.read(4))[0]
         cols = struct.unpack('>I', f.read(4))[0]
         images = np.frombuffer(f.read(), dtype=np.uint8, count = num * rows * cols)
-        # split them up into an array of flat image vectors, so that first axis corresponds to labels
-        images = images.reshape(num, rows * cols)
+        # split them up into an array of flat image column vectors
+        images = images.reshape(num, rows * cols).T
 
     with open(labels_file, 'rb') as f:
         # validate magic
diff --git a/read_display_mnist.py b/read_display_mnist.py
index fd85ee3..3030a0b 100755
--- a/read_display_mnist.py
+++ b/read_display_mnist.py
@@ -10,5 +10,5 @@ train_images, train_labels, rows, cols = mnist.load('train-images-idx3-ubyte', '
 # show the first bunch of training data
 for i in range(10):
     print(f'train image #{i}: label {train_labels[i]}')
-    plt.imshow(train_images[i].reshape(rows, cols), cmap='gray')
+    plt.imshow(train_images[:, i].reshape(rows, cols), cmap='gray')
     plt.show()