Public | Automated Build

Last pushed: 9 months ago
Short Description
An image that can be used to quickly start a jupyter notebook with tensorflow, setup with anaconda.
Full Description

Docker - Anaconda and Tensorflow

Welcome to the anaconda/tensorflow image. This image was built to be used as a simple way
to run a standard ML system with tensorflow.

Structure

The latest branch will (by default) be setup to use the latest version of python and
tensorflow that are CPU based. This will allow for the system to use the standard
ubuntu base image.

However there are tags that will be created for the gpu specific versions of the system.
These will each be tagged as such.

So for every tensorflow example that exists a corresponding version will also exist that
will just have the appended tag value of -gpu, for example:

    latest           ->    latest-gpu 
    python35         ->    python35-gpu
    python35-onbuild ->    python35-onbuild-gpu

Running

To use this image you will want to have a volume mounted at /notebooks, and expose port

  1. docker run -it --rm -p 8888:8888 -v $PWD:/notebooks mikewright/anaconda-tensorflow:latest-gpu

At this point you can open your browser to localhost:8888.

Building

If you want to have your own tools installed you may do so by creating a Dockerfile based on the
onbuild tag. You will need to have your anaconda environment file created and named environment.yml,
and you will want to set the environment variable CONDA_ENV to the name of the enviroment you created.

    -- environment.yml
    name: myenv
    dependencies:
      - jupyter
      - pymc 

    -- Dockerfile
    FROM mikewright/anaconda-tensorflow:latest-onbuild

    ENV CONDA_ENV myenv
Docker Pull Command
Owner
mikewright

Comments (3)
inksci
2 months ago

First,I run
sudo docker run -it --rm -p 9003:8888 -v $PWD:/notebooks mikewright/anaconda-tensorflow:python35-gpu
Later, open
http://0.0.0.0:9003/
within browser.
And click 'new', there are 3 types of python on list.
The sad news is I always get such failure:

Creating new notebook in /apple-ink
[I 15:07:25.990 NotebookApp] Writing notebook-signing key to /root/.local/share/jupyter/notebook_secret
[I 15:07:27.406 NotebookApp] Kernel started: 28fa021e-67b7-4eba-9ffe-acccffbfc7e5
[I 15:07:30.405 NotebookApp] KernelRestarter: restarting kernel (1/5)
[I 15:07:33.415 NotebookApp] KernelRestarter: restarting kernel (2/5)
[I 15:07:36.424 NotebookApp] KernelRestarter: restarting kernel (3/5)
[W 15:07:37.609 NotebookApp] Timeout waiting for kernel_info reply from 28fa021e-67b7-4eba-9ffe-acccffbfc7e5
[I 15:07:39.433 NotebookApp] KernelRestarter: restarting kernel (4/5)
WARNING:root:kernel 28fa021e-67b7-4eba-9ffe-acccffbfc7e5 restarted
[W 15:07:42.441 NotebookApp] KernelRestarter: restart failed
[W 15:07:42.441 NotebookApp] Kernel 28fa021e-67b7-4eba-9ffe-acccffbfc7e5 died, removing from map.
ERROR:root:kernel 28fa021e-67b7-4eba-9ffe-acccffbfc7e5 restarted failed!
[W 15:07:42.478 NotebookApp] Kernel deleted before session
[W 15:07:42.479 NotebookApp] 410 DELETE /api/sessions/57a1d665-9788-4bfc-aeea-910b42ba4e8d (172.17.0.1) 1.60ms referer=http://0.0.0.0:9003/notebooks/apple-ink/Untitled.ipynb?kernel_name=conda-env-tensorflow-py

vigneshj6
8 months ago

@muneeb2405
Hi,
click the link in the terminal which has the token
example :
[I 12:22:57.337 NotebookApp] The Jupyter Notebook is running at: http://0.0.0.0:8888/?token=7edfc3db8fbd9b40910a6efaa00e8e047a2615fcbd285179

muneeb2405
9 months ago

Hi,
when i started the jupyter ,it wanted a password?