/MNIST-CNN-batch-norm

Classification MNIST data using batch normalized CNN

Primary LanguagePythonMIT LicenseMIT

MNIST_CNN_batch_norm

Classification of MNIST_data using CNN or batch normalized CNN

Train CNN with TF-slim which can batch normalization

Usage

Command

python main.py cnn_normal or python main.py cnn_batch_norm

Arguments

Required

  • select_net : select cnn_norm or cnn_batch_norm

Optional

  • --activation_func : choice(sigmoid, relu, lrelu). Default : relu
  • --feature_map1 : Number of feature map1. Default : 64
  • --feature_map2 : Number of feature map2. Default : 128
  • --feature_map3 : Number of feature map3. Default : 256
  • --filter_size : size of filter. Default : 3
  • --pool_size : size of max_pooling. Default : 2
  • --epoch : Number of epochs to run. Default : 10
  • --batch_size : Number of batch_size to run. Default : 100
  • --learning_rate : Learning rate for Adam optimizer. Default : 0.001
  • --drop_rate : Prob of dropout. Default : 0.7

Results

python main.py cnn_batch_norm

result

Reference Implementations