The Keras tutorials by leriomaggio is a great place to start learning Keras. It comes with structured Jupyter notebooks teaching Keras basics all the way to advanced concepts.
You can start by cloning the project repository and initializing a floyd project.
git clone https://github.com/leriomaggio/deep-learning-keras-tensorflow
cd deep-learning-keras-tensorflow
floyd init deep-learning-keras-tensorflow
If you don't have floyd already installed, you can run:
pip install -U floyd-cli
floyd login
You can run this project on a GPU instance on Floyd with a single command:
floyd run --env keras --gpu --mode jupyter
The output should contain the url to the jupyter notebook:
Path to jupyter notebook: https://www.floydhub.com:8000/2sN82HNJhwsqUm5QbWtqUi
You can now open the Jupyter notebook url in your browser and start running the notebooks.
If you want to run this on a CPU instance:
floyd run --env keras --mode jupyter
And if you want a Python 2 environment, you can change it in the command line:
floyd run --env keras:py2 --gpu --mode jupyter
The full list of supported enviroments is here.