Welcome to the documentation for NEANIAS AI Science Gateway: service for Development of Machine Learning Model using Jupyter Hub

About

A web-based development environment for scientists to design, develop and evaluate machine learning solutions. The core technology is Python, which is the most popular programming language in machine learning and data science scenes. Jupyter Hub is a multi-user development environment using the IPython interface, allowing code implementation and interpretation using a web interface. The kernel used includes TensorFlow and Keras (www.keras.io), as the most popular libraries for neural network development, and other relevant packages as well.

Endpoint

The service is available at http://90.147.190.27 .

Access

The service is using NEANIAS Identity Provider for authentication.

Usage

After successful login, Jupyter notebook is automatically launched. In this notebook, Python 3 programming language is supported. The following libraries are preinstalled:

  • tensorflow
  • numpy==1.18.5
  • matplotlib
  • dask
  • scikit-image
  • Keras
  • imgaug
  • scikit-learn
  • astropy
  • xarray
  • rasterio
  • geopandas
  • pyproj
  • jupyterlab

Documentation on Jupyter can be found at https://jupyter.org/documentation .

There is possibility to mount remote storage. Launch a new terminal and run the following:

mkdir $HOME/mystorage
webdav_mount.sh <webdav_url_of_your_storage> $HOME/mystorage <username> <password>

In case of NEANIAS NextCloud ( https://files.dev.neanias.eu ), use https://files.dev.neanias.eu/remote.php/dav/files/<username> as webdav url. Username and password can be generated under User settings/Security/Devices & sessions.

Unmounting a remote storage can be performed by running the webdav_umount.sh script without arguments.

Contact

Please, contact Jozsef Kovacs ( jozsef dot kovacs at sztaki dot hu ) for any assistance.