Skip to content

Training a sequential model to classify handwritten digits by using the mnist dataset and creating a interface to classify your own handwritten digits.

Notifications You must be signed in to change notification settings

leondorian/handwritten_digit_classifier

Repository files navigation

handwritten digit classifier

model_training.py:

Using the mnist dataset to train a model.

model_evaluation.py:

Evaluation of the trained model.

Visualization:
digit_8_gif

main.py:

Creating a interface so we can draw digits that the model then tries to predict.
The canvas to draw on has a size of 280x280, but the image is scaled down to 28x28 to fit the input-size of the model.

Example

digit_8_gif

While drawing the interface shows the current prediction and its confidence.

Pred: 1 (99.43% confidence) Pred: 8 (99.92% confidence) Pred: 7 (98.06% confidence)
digit_1 digit_8 digit_7

About

Training a sequential model to classify handwritten digits by using the mnist dataset and creating a interface to classify your own handwritten digits.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages