Use Julia on Google Colab

date

slug

status

summary

type

tags

Google Colab by default supports Python 3, which can not be directly altered. But it do support running shell command in the cells, and in terminal if you have Pro subscription. This allows us to change the default kernel.

This discussion explains how to use Julia on Colab:

使用google colab实现Julia GPU加速

这篇文章的主要内容来自于 https://discourse.julialang.org/t/julia-on-google-colab-free-gpu-accelerated-shareable-notebooks/15319 本文不过是进行了翻译、整理和一些修改，希望对大家有所帮助。无法正常访问谷歌的可以忽略此帖子。在本地创建 test.txt 文件，写入以下内容 { "nbformat": 4, "nbformat_minor": 0, "metadata": { "colab": { "name": "Julia on Colab.ipynb", "version": "0.3.2", "provenance": [] }, "kernelspec": { "name": "julia-1.4", "display_name": "Julia 1.4" }, "accelerator": "GPU" }, "cells": [ { ...

https://discourse.juliacn.com/t/topic/3239

TL;DR — I’ve made a template notebook. Open this link:

Google Colaboratory

https://colab.research.google.com/drive/1Up_-O2Sp0BVdxBXPhrkwV8bTPtFklH2U?usp=sharing

Copy the notebook to your Google Drive. Anytime you need a Julia kernel, just copy the template to a new notebook, open it, and run the first cell. It will take 3~5 minutes to install the dependencies. Then refresh the page and the Julia kernel is good to go.