Skip to content

LEAP Jupyter Hub

LEAP's primary data and computational resources are available through the LEAP Jupyter Hub. The Jupyter Hub is a shared "cloud-based" computing environment running on cloud-based storage and compute resources.

The hub is designed primarily around interactive use, however long running jobs are also possible.

To gain access to the Hub please see the registration page. LEAP's JupyterHub is managed by our partner 2i2c.

The Hub includes

  • computing resources accessible via a web browser or VSCode as described on this page. The resources include a variety of hardware configurations and a range of pre-defined or custom software environments.

  • data resources divided among generous shared "cloud buckets" for data and limited home directories for scripts etc. as described at Where Data Lives.

The computing resources have fast access to LEAP's data resources; they also have fast connections to the broader internet which lowers the barrier to working with data held elsewhere.

Compared to using:

  • a laptop, the Hub offers fast access to data and the ability to access much more powerful computational resources including GPUs for ML training tasks.
  • HPCs, clusters, etc. the hub offers simpler access and less competition for resources (since the pool from which the resources are drawn is enormous). It does not, however, easily support non-interactive use.

Moving data in to data resources is free; moving data out is expensive (see Data Lifecycle). Please plan your workflow accordingly.