Skip to content

Latest commit

 

History

History
76 lines (44 loc) · 2.54 KB

README.md

File metadata and controls

76 lines (44 loc) · 2.54 KB
Raw

IML 2019 Tutorials

The tutorials run in environments with either TensorFlow v1 or v2:

  • Introduction tutorial: TensorFlow v1
  • Feature engineering tutorial: TensorFlow v2
  • GAN tutorial: TensorFlow v1

Be aware of slight differences in their setup as explained in the following.

In general, there are four ways to start the notebooks.

1. SWAN

Click on the following link:

SWAN

You will be asked to configure the environment in a small dialog.

  • For the introduction and the GAN tutorial, enter the following environment script path and then press the "Start my Session" button at the bottom:
/eos/user/m/mrieger/public/iml2019/intro/setup.sh
  • For the feature engineering tutorial, do the same with the following environment script path:
/eos/user/m/mrieger/public/iml2019/lbn/setup.sh

2. Binder

For TensorFlow v1:

Binder

For TensorFlow v2:

Binder

3. Standalone docker image from the docker hub

Make sure not to execute the following commands with sudo as a port will be opened on your machine to run and host the notebook server. Otherwise, you potentially allow people within you local network to access your system with root permissions!

If you don't have the permission to execute docker with your user account, add yourself to the "docker" group (e.g. via sudo usermod -a -G docker $(whoami)).

# for TensorFlow v1 (introduction and GAN tutorials)
docker run -ti -u $(id -u):$(id -g) -p 8888:8888 3pia/iml2019:tf1

# for TensorFlow v2 (feature engineering tutorial)
docker run -ti -u $(id -u):$(id -g) -p 8888:8888 3pia/iml2019:tf2

4. Docker image with a local checkout

You can also check out the repository and use the script located in docker/tf{1,2}/run.sh to start the docker container. The script will run the same command as above and mounts the repository into the container. This way, changes you make in the notebooks are persistently stored within you local checkout.

As above, make sure not to run the container as root!

git clone https://github.com/3pia/iml2019
cd iml2019

# for TensorFlow v1 (introduction and GAN tutorials)
./docker/tf1/run.sh

# for TensorFlow v2 (feature engineering tutorial)
./docker/tf2/run.sh