sgudbrandsson / tensorflow-keras-docker

Tensorflow + Keras as a Docker container for local experiments

Docker image for running Keras with a Tensorflow backend

This Dockerfile builds an image containing both Tensorflow and Theano, but enables TF as a backend for Keras.

You can add any files you like to this project and rebuild the image to experiment with them.

Usage

On OSX, you will need to install Docker Machine or boot2docker before running this image.

Building the image:

docker build . --tag peterrosell/keras-jupyter

Running:

1. Make a directory named notebooks in your home directory.

2. Run the following command from command line

docker run \
	-it \
	--rm \
	--name keras \
	-e KERAS_BACKEND=tensorflow \
	-p 8888:8888 \
	-p 6006:6006 \
	-v ${HOME}/notebooks:/notebooks \
	peterrosell/keras-jupyter

3. Browse to http://localhost:8888 to access the GUI for jupyter Notebooks. Browse to http://localhost:6006 to access Tensorboard. Tensorboard is used for displaying log results from a learing session.

4. To shutdown just press Ctrl-C at command line.