Google is offering free TPU and GPU for AI using Colaboratory (Colab)

Google anounced their new Colaboratory (colab), which is a free Jupyter notebook environment that requires no setup runs entirely in the cloud.

With Colaboratory you can write and execute code, save and share your analyses, and access powerful computing resources, all for free from your browser.

But the interasting thing is that it’s free!

We don’t need to pay for TPU or GPU processing even Google will provide their own RAM and CPU. If are looking for a high end machine or cloud server, now you don’t need to spend a lots of money over PC buding or server fees. You can compile and execute your code by using cloud resources. You would also get some tutorials along with that. We can train Machine Learning and Deep Learning Model with the help of GPU and TPU.

TPU Memory Limit in Google Colaboratory or Google Colab

Connected to "Python 3 Google Compute Engine Backend (TPU v2)"
TPU: 64GB
RAM: 0.66GB / 12.72GB
Disk: 20.29GB / 48.97GB

GPU Memory Limit in Google Colaboratory or Google Colab

Connected to "Python 3 Google Compute Engine Backend (GPU)"
GPU: 121MB / 11.44GB
RAM: 1.07GB / 12.72GB
Disk: 22GB / 358.27GB

What are the charges for cloud server for AI?

Google is helping us to create a beautiful future using AI and Natural Language Processing. Every student or researchercan’t afford a GPU based PC or can’t pay around 5 $ dollar per hour for a cloud server. AWS (amazon web server) and Google Cloud Sever are providing GPU based cloud processing aroud 5 $ dollar per hour.

World’s biggest press company (Associated Press) is using AI to increase their business?

Now everyone is talking about AI, Deep Learning, Machine Learning and Natural Language Processing, because according to a report world’s 80% data is generated within last two year and it’s evolving day by day.
So many e-commerce sites, social sites and IT companies are using AI for generating large revenue like (Associated Press, which is the world’s best press. AP increase their estimate value around 3-4% by AI)

These companies can pay for cloud server to run their business, but if we want to start a new business and take AI for smart features, so we can go with Google Colaboratory Research Program. Where we have two options first one is GPU (Graphical Processiing Unit) processing and second one is TPU (Tensor Processing Unit) processing.

What is GPU (Graphical Processing Unit)?

We already know about Graphical Processing Unit, which is a processing unit for graphics to perform complex and image process. It performs that work which a CPU takes a lot of time. It use to play video games, video editing, Bitcoin mining and graphics related work.

What is TPU (Tensor Processing Unit)?

Now talk about TPU (Tensor Processing Unit), It’s not new. TPU introduced first by Google in I/O 2017, now TPU has two versions (2.0 and 3.0). In 2018 in Google I/O anounced that they are using liquid cooling in their TPU hardware.
TPU helps in that field where GPU become fail. TPU being use in Machine Learning, Deep Learning and AI (Artificial Intelligence), because it’s design to performe AI based instruction. TPU uses some AI features for processing and with the help of this smart working it become a fastest processing device. We can say that TPU = GPU + AI

TPU is more expensive device than a GPU. But you can use it’s processing by paying around 6.38 $ per hours to Google cloud.