Added all the stuff from part 3

except for the categorical features I had type 1 and type 2

Instead of one hot encoding
I mapped each typing to a number and changed the columns values

I also dropped name and generation

Then I used leaky relu and early stopping
+ more epoch

While there wasnt overfitting present
the model now preforms better with the saved model getting 96% on testing

All coding done then I will push reports