TensorFlow Dockerfiles

Base on: https://github.com/tensorflow/tensorflow/tree/master/tensorflow/tools/dockerfiles
See ORI_README.md for more detail

Building

Build from dockerfile to create image, tf is tag.

$ docker build -f ./dockerfiles/cpu.Dockerfile -t tf .

Running

Run with -v to shared directory from your local directory.

# container with tensorboard run on port 6006, and needs a volume for workdir
$ docker run \
  --name tensorboard \
  --user $(id -u):$(id -g) \
  -p 6006:6006 \
  -v $(pwd):/workdir \
  -it tf \
  tensorboard --logdir workdir/pyDir

# Place event file under workdir/pyDir to display on tensorboard.
# On Mac I test $(pwd)/tsboardDir but always display
# "No dashboards are active for the current data set." on browser.
# Works well after remove "$(pwd)"

# container with bash
$ docker run --name bashCmd -v $(pwd):/workdir -it tf /bin/bash

# cmd1: container with jupyter
$ docker run --user $(id -u):$(id -g) -p 8888:8888 -v $(pwd):/notebooks -it tf
# cmd2: cd to notebooks dir
$ cd /notebooks
# cmd2: run following cmd in container to start jupyter
$ ./tools/run_jupyter.sh
# Launch browser, link with: http://localhost:8888/ and input token (after run cmd1 you can get a token).

Edit Files

On your local side install software to edit .py files. Such as Visual Studio Code.

Build .py

On bashCmd container terminal run floowing command to build .py

# run cmd ex. python test.py
$ python YOUR_PY_FILE_NAME

Tensorboard

After Create "evnet" file by run .py
(you need to write some code about tensorboard in .py to create event file.)
Launch browser, link with: http://localhost:6006/

Start/Stop containers

$ docker stop/start tensorboard

Known Issue

Jupyter: In this dockerfile, run on Mac platform cannot display event file on tensorboard and cannot display running plot. Current cannot find solution. Add "%matplotlib inline" on the top of .py file can display plot but cannot running.

Jupyter Cmd

'Esc + d' twice to remove blank line.

Docker Cmd

# In ./tools/cleanup.sh
# Stop all containers
$ docker stop $(docker ps -a -q)

# Remove all containers
$ docker rm $(docker ps -a -q)

# Remove all images
$ docker rmi $(docker images -a -q)

Spark Image build

$ docker build -f ./dockerfiles/cpu-spark.Dockerfile -t spark .

Spark Cmd

# container with bash
$ cd ${your_workdir}
$ docker run --name sparkCmd -v $(pwd):/workdir -it spark /bin/bash