lamnguyenkhoa/mnist-practice

A repo used to study classification of handwritten characters

Handwriting digits and letters recognition

All trained models are placed in /trained_models. Current best model is homemade_model4

Put all data files into /training_data in root folder

Dataset used in this project:

• EMNIST-bymerge, EMNIST-letter dataset: https://www.kaggle.com/crawford/emnist
• A-Z handwritten alphabets: https://www.kaggle.com/sachinpatel21/az-handwritten-alphabets-in-csv-format
• MNIST digits 0-9 dataset from keras (built-in)

References

• https://machinelearningmastery.com/how-to-configure-image-data-augmentation-when-training-deep-learning-neural-networks/

• https://machinelearningmastery.com/how-to-implement-major-architecture-innovations-for-convolutional-neural-networks/