lopezco/docker-tensorflow

Running the Docker container from the Dockerhub image

docker run -p 8888:8888 -p 6006:6006 --name udacity-tensorflow -it lopezco/udacity-tensorflow

Note that if you ever exit the container, you can return to it using:

docker start -ai udacity-tensorflow

Environment Variables

You can set some environment variables while running the container using the --env parameter as follows:

--env JUPYTER_PASSWORD="my_password" --env JUPYTER_PORT=8888 --env TENSORBOARD_LOGDIR="./logs" --env TENSORBOARD_PORT=6006

Accessing the Notebooks

• Jupyter: http://localhost:8888
• Tensorboard: http://localhost:6006

Running the Docker container with docker-compose

Install docker-compose following this guide. Then, download this docker-compose file.

docker-compose up [-d]

This command will pull the image from docker-hub and automatically run the container. The parameter -d is optional (run in detached mode)

Note that if you ever stop the container, you can restart it using:

docker-compose up [-d]

Environment Variables

You can set some environment variables in the docker-compose file to control the container:

environment:
  JUPYTER_PASSWORD: "my_password"
  JUPYTER_PORT: 8899
  TENSORBOARD_LOGDIR: ./logs
  TENSORBOARD_PORT: 8008

Accessing the Notebooks

• Jupyter: http://localhost:8899
• Tensorboard: http://localhost:8008

## History of base image

Base image: gcr.io/tensorflow/tensorflow:latest

• 0.1.0: Initial release.
• 0.2.0: Many fixes, including lower memory footprint and support for Python 3.
• 0.3.0: Use 0.7.1 release.
• 0.4.0: Move notMMNIST data for Google Cloud.
• 0.5.0: Actually use 0.7.1 release.
• 0.6.0: Update to TF 0.10.0, add libjpeg (for Pillow).
• 1.0.0: Update to TF 1.0.0 release.
• 1.1.0: Update to TF 1.3.0 release.