any plans for Julia?

[dylanmmarshall]

Sorry if not the place for the question but after all, CoLab was modeled after Jupyter and Jupyter stands for Julia / Python / R... Am wondering if there are plans for Julia use in CoLab or not?

[ViralBShah]

I wonder what the result of the triage was. We in the Julia community are happy to collaborate and help make it happen.

[dylanmmarshall]

Would be nice as it would mean I wouldn't have to leave my favorite tool 🤗

[Oblynx]

There are some instructions going around for running Julia on Colab in a rather hacky way, also with GPUs:

• https://dev117uday.medium.com/how-to-run-julia-on-google-colab-dd631e6e7e43
• https://discourse.julialang.org/t/julia-on-google-colab-free-gpu-accelerated-shareable-notebooks/15319

I tried to follow them but was unsuccessful. Maybe things have changed on the Colab side since these hints came out; after all they're pretty hacky.
In the meantime I'm giving https://nextjournal.com/ a try 😄

[oschulz]

Now that Julia has matured even more, maybe this could be revisited?

There are (hacky) ways of running Julia on Colab, but they are not user friendly. They demonstrate that it shouldn't be a problem technically at all though - Colab basically would just need to provide Julia binaries and preinstall the "IJulia.jl" package.

[Dsantra92]

Seems like the hacky methods are no longer working. Running any julia command using ! results in error. Tried this with multiple Julia versions (1.7+ and 1.8+ s).

!julia -e "a=10"

Results in

munmap_chunk(): invalid pointer

signal (6): Aborted
in expression starting at none:0
gsignal at /lib/x86_64-linux-gnu/libc.so.6 (unknown line)
abort at /lib/x86_64-linux-gnu/libc.so.6 (unknown line)
unknown function (ip: 0x7fde015d926d)
unknown function (ip: 0x7fde015e12fb)
unknown function (ip: 0x7fde015e154b)
close_unit_1 at /workspace/srcdir/gcc-11.1.0/libgfortran/io/unit.c:742
close_units at /workspace/srcdir/gcc-11.1.0/libgfortran/io/unit.c:800
unknown function (ip: 0x7fde0198ef6a)
unknown function (ip: 0x7fde015928a6)
exit at /lib/x86_64-linux-gnu/libc.so.6 (unknown line)
main at julia (unknown line)
__libc_start_main at /lib/x86_64-linux-gnu/libc.so.6 (unknown line)
unknown function (ip: 0x400808)
Allocations: 2722 (Pool: 2712; Big: 10); GC: 0

[SoumenMaity922]

Now I am Julia 1.8.5 in my local machine, but
I cannot able to use Julia in google Collab . In the runtime type, there is only one option to change the hardware accelerator. So, how can I use Julia in collab

[metrizable]

As commented on #3385 and #3391, installing and using julia in Colab should now be working.

[oschulz]

Now if Julia were just included in Colab by default ... :-)

[MilesCranmer]

Confirmed everything is working thanks to @metrizable and the colab's team help in #3385.

For running Julia, check out @ageron's https://github.com/ageron/julia_notebooks for an example of replacing the Python kernel with Julia, so you can run Julia code directly inside the notebook.

[oschulz]

The reload and so on isn't really beginner/classroom-friendly, though. Any chance that Julia could be included in Colab?