installing requirements: conda install -c conda-forge mnist
sample python code for accessing the dataset :
from mnist import MNIST
mndata = MNIST(r'C:\Users\krish\OneDrive\Documents\krishna\sem6\DL\mnist')
mndata.gz = True
train_images = np.zeros([1000,28,28])
test_images = np.zeros([1000,28,28])
images_tr,train_labels = mndata.load_training()
images_tr = np.array(images_tr[:1000])
train_labels = np.array(train_labels[:1000])
images_ts1,test_labels = mndata.load_testing()
images_ts = np.array(images_ts1[:1000])
test_labels = np.array(test_labels[0:1000])
for i in range(1000):
train_images[i] = images_tr[i].reshape(28,28)
test_images[i] = images_ts[i].reshape(28,28)