An introduction to tensorflow via keras
This will be much easier than the local installation most of the time. The first step is to make an account over at Floydhub. You will then want to create a project (see here for more details). I would suggest that you name it: tf-keras-tutorial
.
Once that is done you will need to download this repository:
git clone https://github.com/knathanieltucker/tf-keras-tutorial.git
cd tf-keras-tutorial
Install the floydhub cli:
pip install floyd-cli
Initialize the project:
floyd init tf-keras-tutorial
This may prompt you to login. And finally run:
floyd run --cpu --env keras:py2 --tensorboard --mode jupyter
And that should take you to a working notebook. At this point I would suggest you turn the job off and wait for class to turn it back on again!
Follow guide here to set up basic virtual env. After, run:
pip install -r requirements.txt
to install all requirements.
Make sure that you can run ipython notebook
in your virtual env.