Example of Image Recognition using tensorflow/tflearn

Setup and installation of environment for osx. Will add information for linux soon. More details on setting up anaconda can be found here: http://machinelearningmastery.com/setup-python-environment-machine-learning-deep-learning-anaconda/

ANACONDA 4.3.1

Download and install Anaconda:

wget https://repo.continuum.io/archive/$ANACONDA_NAME
bash $ANACONDA_NAME

source bash_profile

source ~/.bash_profile

Create environment and install conda-build

conda create --name ml python=3.6
conda install conda-build

Activate "ml" environment before installing packages.

source activate ml

Your terminal should look like this:

(ml) computer_name/folder$

Install python packages

pip install jupyter
pip install scikit-learn
pip install matplotlib
pip install tflearn
pip install pillow
pip install scipy

Install OPENCV 3.2.0

brew install git cmake pkg-config jpeg libpng libtiff openexr eigen tbb
pip install opencv-python

Install Tensorflow 1.0.1

TF_URL=https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-1.0.1-py3-none-any.whl
pip install --ignore-installed --upgrade $TF_URL

Make environment available in Jupyter Notebook

pip install ipykernel
python -m ipykernel install --user --name ml --display-name "Python (ml)"

Start notebook

jupyter notebook

Remember to set Kernel to "Python(ml)" when you open a notebook.