/Basic-Deep-Neural-Networks-Applied-on-MNIST-Handwritten-Digit-Dataset

Primary LanguagePython

Basic Deep Neural Networks Applied on MNIST Handwritting Dataset

1/ Hierarchy:

  • 01_dnn_minist : MPL architecture

    • network : Learning model description
    • step02_training.py : Training & Testing Process
    • run.sh : Bash script to call 'step02_training.py' ---> Other architectures show similar files

  • 02_cnn_mnist: CNN architecture

  • 03_cnn_mixup_minist: CNN architecture with mixup data augmentation

  • 04_rnn_mnist: RNN architecture (Bidirectional RNN with GRU cell)

  • 05_crnn_mnist: CRNN architecture

  • 06_cnn_rnn_parallell: Parallel CNN & RNN architecture

  • README.md

2/ Notes:

  • All networks base on tensorflow framework 1.X
  • Dataset : MNIST