This Supervised model uses a small part of Original MNIST dataset that had 60,000 training samples, and 10,000 test samples. Each image is represented by 28x28 pixels, each having a grayscale value of 0 - 255
here it uses a simplified version of training the model upon 0-9,+,-,*,/ only to fetch the answer of basic handwritten equation. This idea has possibility of wide variety of features that could be added to it and deployed
If you fork/copy this code. please do send me what improvements did you make to the code as that would make me very happy.
-
Downoad the dataset and extract the necessary files Link to download the dataset: https://drive.google.com/file/d/1Rz0V529mT-hz2eiu8O-S1ZINk3ZotG4w/view?usp=sharing
-
Train the CNN and SAVE THE MODEL
-
Run the Equation notebook to obtain the output
(Runs better with GPU timings)