Google Colab Notebooks

 

Hmmm... Imagine this. You made a program, A great Machine Learning program that can go on possibly forever, requiring a whole lot of resources, yet you celebrated your computer's fifth anniversary last summer.(Why do I sound like an advertisement?) Now, you bewitch the minds and ensnare the senses of the people patently possessing greatest processing power and use their computers instead. Ostentatious, right? Perhaps you won't need sorcery.

Google's Colaboratory is a frontend User Interface(UI) for Cloud Computing. Cloud Computing, as you may know, is way of acquiring greater computer resources over the web, especially Cloud Storage and Computing Power which would have otherwise taken a lot of effort and capital to set up. Usually cloud computing is subscription based service, However some providers give it for free or as a trial with limited period and/or resources, which is the case with Google Colaboratory.

Origins

So back in 2014, Fernando Pérez announced a spin-off project from IPython called Project Jupyter. As soon it got released it immediately gained popularity. This is because of three primary reasons — One, It's support for multiple languages including Julia, Python and R, Second — That it can work on your favourite browser, and Third — It's cell based REPL structure. It was soon adopted as frontend for cloud computing by many providers due to its ease of use and browser-based UI, and so did Google, which itself sponsored Project Jupyter.

What colab offers if you're a developer?

• Write and execute code in Python
• Document and illustrate your code (that supports MathJax)
• Create/Upload/Share notebooks
• Import/Save notebooks from/to Google Drive
• Free Cloud service with free GPU (you can go for Colab Pro if you want to...)

But I'm NOT a developer!!!

There's nothing you've got to worry about. Colab can prove useful for you even if you're a normal user knowing just a wee bit of coding (just to let you know I advocate Allman's with all my heart). Be it efficiently managing your google drive, Transferring files from one drive to another, zipping files in the cloud before you download/upload them, downloading files directly to your drive without being counterproductive by downloading it to your computer and then uploading it, OR, making it act like a proxy before you download anything from anywhere (even the greatest secretest servers/IPs blocked or censored wouldn't deny google's IP from accessing it, as it is OF no particular country), OR, doing your regular math homework, Calculating equations, Plotting graphs (although that requires a little bit more exp), Colab can help you do it all, even more. 

If you look at it, you can even stitch together multliple drive spaces and tadaa... you have space in multiples of Fifteen GeeBee surpassing that limit on drive space that can support practically nothing in the modern world, for free (Not "that" easy though...) 

Is it all that Serene?

No. No. There is a basic limit on resources and time. Google Colab has a ‘maximum lifetime’ limit of running notebooks that is 12 hours with the browser open, and the ‘Idle’ notebook instance is interrupted after 90 minutes (Which can be tampered with😏). In addition, you can have a maximum of 2 notebooks running simultaneously. As for CPUs, GPUs, Memory and disk space, We get a Intel(R) Xeon(R) CPU @ 2.20GHz CPU, NVIDIA(R) Tesla(R) K80 GPU, an approximate 12GB of Memory which can be expanded to 25 GB, and around 100GB Disk space which has ≈60GB free space. Colab have dynamic usage limits that sometimes fluctuate. This means that overall usage limits as well as idle timeout periods, maximum VM lifetime, GPU types available, and other factors vary over time. They can (and sometimes do) vary quickly.

And as for second; if you're thinking that you've got the perfect way to do illegal online stuff, beware! This is a clear infringement of their policy, and you can get your account suspended, or worse, end up facing a trial!