Deep Learning Model.md

1. CNN

• http://khanrc.tistory.com/entry/TensorFlow-5-MNIST-CNN
  
• http://pythonkim.tistory.com/56
  
• http://khanrc.tistory.com/entry/TensorFlow-5-MNIST-CNN
  
• http://cs231n.github.io/understanding-cnn/
  
  

- Google Inception Model

• https://norman3.github.io/papers/docs/google_inception.html
  

  • Inception V3
    

• http://solarisailab.com/archives/1422
  

• Achieving 90% accuracy in Object Recognition Task on CIFAR-10 Dataset with Keras: Convolutional Neural Networks
  

• MNIST Image Class. TensorFlow CNN 99.51% Test Acc (18/12/30)
  

2. RNN

https://www.tensorflow.org/tutorials/sequences/recurrent

**- vanilla rnn**
http://www.easy-tensorflow.com/tf-tutorials/recurrent-neural-networks/vanilla-rnn-for-classification
**- LSTM**
http://www.hanbit.co.kr/media/channel/view.html?cms_code=CMS6074576268
**-image captioning**
https://github.com/tensorflow/models/tree/master/research/im2txt
-keras : Repeactvector

Normalization Technique

- batch norm
http://openresearch.ai/t/topic/80
- dropout problem
https://stackoverflow.com/questions/39691902/ordering-of-batch-normalization-and-dropout-in-tensorflow