Installing Julia, and a basic handful of machine learning packages, takes 20 minutes in Colab (and other REPL environments). This notebook loads precompiled binaries and cuts the installation time to ~30 seconds. Just hit "Open in Colab" above, run the initialization cell, and start working.

This downloads a Julia binary (from julia.org) and a working .julia "depot" directory (from this repo). Then it installs IJulia for use in the notebook. The depot directory is tarballed with compile.sh. Ideally, we'd use PackageCompiler to produce a sysimage; unfortunately, as of this writing, this does not successfully copy over artifacts / JLLs. This notebook is not meant to be a long-term solution - just something to get you up and running on Colab.

The source code in this directory is distributed under the MIT license. The binaries in depot/, however, are redistributed under their original terms. These may include e.g. GPL-licensed software.