Setup

Downloading and installing data science-related python packages manually can be hard and tedious. Anaconda to the rescue! It's basically a bundle that includes python and a number of common python dependencies. It may take a few minutes to download and install.

Windows

Download and install Anaconda (Python 3.6 version, 64-bit installer): https://www.continuum.io/downloads#windows

Installation notes:

• If you already have Python installed on your system, it is wise to uncheck "Add Anaconda to my PATH environment variable" and "Register Anaconda as my default python 3.6". The following instructions assume that these two options have been unchecked.
• You probably don't need Anaconda Cloud

Start "Anaconda Prompt" and run

• conda create --name tensorflow python=3.5
• activate tensorflow
• conda install scipy
• pip install tensorflow==1.0.1
• pip install Keras==2.0.2
• conda install scikit-image

Check your installation by running python -c "import keras". If no error appears, the installation should be fine. If it outputs "Using TensorFlow backend.", that's okay.

Ubuntu

https://docs.continuum.io/anaconda/install-linux

• cd ~/anaconda3/bin
• ./conda create --name tensorflow python=3.5
• source activate tensorflow
• Running which python should now print something like /home/yourusername/anaconda3/envs/tensorflow/bin/python
• pip install tensorflow==1.0.1
• pip install Keras==2.0.2
• conda install scikit-image

Check your installation by running python -c "import keras". If no error appears, the installation should be fine. If it outputs "Using TensorFlow backend.", that's okay.

Setting up your IDE

It's a good idea to tell your integrated development environment (IDE) which Python interpreter you're using, so it can run Python scripts in the correct environment and be smart about coding assistance features such as code completion.

More info: https://docs.continuum.io/anaconda/ide_integration

When you're all set up, move on to the tutorial to get started with the programming