CNN (Convolution Neural Networks)

FIRST CONTACT

WITH TENSORFLOW

5. MULTI-LAYER NEURAL

NETWORKS

[참고]

FIRST CONTACT WITH TENSORFLOW

http://www.jorditorres.org/first-contact-with-tensorflow/#cap5

텐서플로우 코리아 (TensorFlow Korea)

https://tensorflowkorea.wordpress.com/5-텐서플로우-다중-레이어-뉴럴-네트워크-first-contact-with-tensorflow/

https://github.com/rickiepark/tfk-notebooks/blob/master/first-contact-with-tensorflow/chapter5_convolution_neural_network.ipynb

TensorFlow Tutorial

https://www.tensorflow.org/versions/r0.9/tutorials/mnist/pros/index.html#convolution-and-pooling

TensorFlow API

https://www.tensorflow.org/versions/r0.9/api_docs/python/nn.html#convolution

모두를 위한 머신러닝/딥러닝 강의

http://hunkim.github.io/ml/

Tensorflow Example 실행

# python –m tensorflow.models.image.mnist.convolutional

http://www.nvidiakoreapsc.com/contest/deeplearning2016.html

Visualizing and Understanding Deep Neural Networks by Matt Zeiler

https://www.youtube.com/watch?v=ghEmQSxT6tw

Chapter 9

Convolutional Networks

http://www.deeplearningbook.org/contents/convnets.html

---

아래 Tutorial 을 보면 Weight 값을 image로 보여주는 Helper Function을 정의했음.

배열값으로 표현된 이미지를 이미지로 보여주는 Helper Function도 정의되어 있음.

* TensorBoard로 Weight 값이 변하는 것을 보여주는 방법은 없을까?  아니면 Weight 값이 변하는 것을 이미지로 보여줄 수 있을까?

* 필터(커널)을 통해서 만들어진 공통 학습모형에서 Label별 Weight값을 적용해서 Label별 학습모형을 볼 수 없을까?

TensorFlow Tutorial #02 - Convolutional Neural Network

https://youtu.be/HMcx-zY8JSg?list=PL9Hr9sNUjfsmEu1ZniY0XpHSzl5uihcXZ

https://github.com/Hvass-Labs/TensorFlow-Tutorials

TensorFlow-Tutorials

Contents

1. Simple Linear Model (Notebook)

2. Convolutional Neural Network (Notebook)

3. PrettyTensor (Notebook)

Videos

These tutorials are also available as YouTube Videos.

Convolutional Neural Network (Notebook) 에서 다음과 같은 에러가 나서 youtube에 질문을 올렸음.

https://www.youtube.com/watch?v=HMcx-zY8JSg&feature=youtu.be&list=PL9Hr9sNUjfsmEu1ZniY0XpHSzl5uihcXZ

TensorFlow - Softmax Regression

http://andersonjo.github.io/machine-learning/2016/05/07/TensorFlow-MNIST/

3. 텐서플로우(TensorFlow)를 이용한 MNIST 문자 인식 프로그램 만들기

http://solarisailab.com/archives/303