How to Install TensorFlow using pip
https://www.tensorflow.org/install/install_linux#determine_how_to_install_tensorflow
Tested on cs50.io:
sudo pip install -U pip
sudo pip install tensorflow
MNIST For ML Beginners
softmax.py from https://www.tensorflow.org/get_started/mnist/beginners
Report Accuracy print.py
TensorBoard Demo
board.py TensorBoard Image image.py
https://www.tensorflow.org/get_started/summaries_and_tensorboard
needs to run in python before tensorboard
python board.py
tensorboard --logdir=/tmp/mnist_logs --port 8080
tensorboard defaut port 6006 may not be open
e.g. https://cicai-eniof.cs50.io
Predict from Image
predict.py
requires :
sudo pip install --upgrade scipy
sudo pip install pillow
based on softmax.py
L25 based on cnnPredict.py L3-11
trained on mnist.train.images[0] vs image imread('test3.png',mode='L')
expert.py from https://www.tensorflow.org/get_started/mnist/pros
Remove Tensorflow Warning
disable "cpu_feature_guard":
export TF_CPP_MIN_LOG_LEVEL=2
2L.py ReLu, Random W
2 Layer Feed-Forward Neural Network
http://stackoverflow.com/questions/38136961/how-to-create-2-layers-neural-network-using-tensorflow-and-python-on-mnist-data
https://cloud.google.com/blog/big-data/2017/01/learn-tensorflow-and-deep-learning-without-a-phd
ply.py on ipynb, view image data. See installJupyter.txt
Original:
https://github.com/random-forests/tutorials/blob/master/ep7.ipynb
https://www.youtube.com/watch?v=Gj0iyo265bc
Updated:
http://tneal.org/post/tensorflow-ipython/TensorFlowMNIST/
http://stackoverflow.com/questions/36651704/which-cmap-colormap-to-use-with-matplotlib-imshow-to-diplay-the-mnist-datase
Predict from Canvas see In [8]:
https://github.com/dmlc/mxnet-notebooks/blob/master/python/tutorials/mnist.ipynb
MNIST for Handwriting Prediction:
https://github.com/niektemme/tensorflow-mnist-predict