knn.py

+import numpy as np
+
+
+def distance_matrix(mat1, mat2):
+    dists = np.sqrt(np.matmul(mat1, mat1)+ np.matmul(mat2, mat2) - 2 * np.matmul(mat1, mat2))
+    return dists
+
+def knn_predict(dists, labels_train, k):
+    return predicted_labels
+
+
+
+
+
+
+
+
+mat1 = np.array([[1, 2],
+                 [3, 4]])
+
+mat2 = np.array([[5, 6],
+                 [7, 8]])
+
+A = np.matmul(mat1, mat1)
+print(A)
+
+B = np.matmul(mat2, mat2)
+print(B)
+
+C = 2 * np.matmul(mat1, mat2)
+print(C)
+
+print(A + B - C)
+
+mat = distance_matrix(mat1, mat2)
+print(mat)
\ No newline at end of file

read_cifar.py

 import pickle
 import numpy as np
 
+
 def unpickle(file):
     with open(file, 'rb') as fo:
         batch = pickle.load(fo, encoding='bytes')
@@ -19,6 +20,11 @@ def read_cifar(folder_path):
         labels = np.concatenate((labels, read_cifar_batch(folder_path + "/data_batch_" + str(i))[1]))
     return data, labels
 
+def split_dataset(data, labels, split):
+    index = int(split * len(data))
+    data_train, data_test = np.split(data, index)
+    labels_train, labels_test = np.split(labels, index)
+    return data_train, labels_train, data_test, labels_test