Running Tensorboard from Shinyproxy

ShinyProxy
#1

I’m looking at running Tensorflow’s Tensorboard from shinyproxy, alongside other shiny apps we have configured.

I’m able to run the container with a docker run command:

docker run -p 6006:6006 tensorflow/tensorflow:2.0.0-py3 tensorboard --logdir /opt/tensorflow --host 0.0.0.0 --port 6006

However when trying to run through shinyproxy I get a container unresponsive error, below is the application.yml extract.

  - id: tensorflow-tensorboard
display-name: Tensorflow Tensorboard
description: A tool for providing the visualizations from the Tensorflow workflow
docker-cmd: ["tensorboard","--log-dir /opt/tensorflow/","--host 0.0.0.0","--port 6006"]
container-image: tensorflow/tensorflow:2.0.0-py3
container-volumes: ["/opt/tensorflow/:/opt/tensorflow/"]
container-env: 
  TZ: Pacific/Auckland  
port: 6006

I’ve tried variations of the cmd/port etc (including using 3838) - is someone able to shed some light on what may be going wrong or your experience of successfully implementing other non-shiny apps in shinyproxy.

Thanks

#2

Hi

I guess it’s a bit a too for you now, but for reference here is a working tensordboard configuration:

  - id: tensorflow-tensorboard
    display-name: Tensorflow Tensorboard
    description: A tool for providing the visualizations from the Tensorflow workflow
    container-cmd: ["tensorboard", "--logdir", "/opt/tensorflow/", "--host", "0.0.0.0", "--port", "6006" ]
    container-image: tensorflow/tensorflow:2.2.3
    container-volumes: [ "/opt/tensorflow/:/opt/tensorflow/" ]
    container-env:
      TZ: Pacific/Auckland
    port: 6006