I trained a simple Recurrent Neural Network which can take string of numbers as input and solve additin or substraction.
Input will be like "55+40" and the model will predict the output 95.
If the input is iike "55-40", model should predict 15.
The hyperparameter are choosen in such a manner so that it can predict 100% correct. There is no dataset given, as I could not find any. So I have created some data using numpy random function.
A custom data is generated with python code. All the source code is provided as a notebook file. I also make seperate files for addition and substraction.