Skip to content

feeney92/Handwritten_Digit_Classifier

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 

Repository files navigation

Digit_Recognition

This code trains a convolutional neural network (CNN) to classify hand written digits using the MNIST data set. The resulting weights of the CNN are saved in h5 format ("CNN_weights_digit_recognition.h5").

The key steps in the model are as follows:

  1. Training and test data sets are created from the MNIST dataset
  2. The training data is augmented by randomly rotating, zooming in, blurring and translating the images
  3. Examples training data images (following augmentation) are shown
  4. The CNN architecture, the optimiser and the learning rate are defined.
  5. The CNN is trained and the resulting weights are saved in h5 format.
  6. The training history of the model is shown.
  7. The accuracy of the model is calculated using the test set.

About

Creating a CNN to classify hand written digits

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages