I assume something like this can be implemented, after going through the dataset link:
import os
import numpy as np
X_train = []
y_train = []
for i in range(10):
dir="DirectoryName/Train/digit_{0}".format(i)
print(i,":",len(os.listdir(dir)))
for j in os.listdir(dir):
X_train.append(plt.imread(dir+"/"+j))
y_train.append(i)
X_train = np.asarray(X_train)
y_train = np.asarray(y_train)
Similarly for the characters directory - train/test
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…