Skip to Content

Amazon SageMaker Studio Lab for Beginners

Posted on October 29, 2022 by

Categories: AWS


Learn how SageMaker Studio Lab, GitHub, and a Kaggle dataset can make studying machine learning (ML) simple and enjoyable.

A lot of things use machine learning. Today, you may choose any article or commercial application to read or use ML. Using a Jupyter notebook is the simplest method to learn about and experiment with machine learning.

The programming language of our choosing, text annotations, and data manipulation are all done in a notebook. Every time I learn anything new, I find myself making quick notes in Evernote, Quip, or Notes. I then compose the final version in a Word or PDF document. ML experimentation ought to be no different. You want to try with ML right away.

Amazon SageMaker Studio Lab: What is it?

A free, no-setup notebook environment created specifically for learning and ML experimentation is called Amazon SageMaker Studio Lab. If you are accustomed to utilising Jupyter notebooks, SageMaker Studio Lab will be a cinch for you. The best thing is that neither an AWS account nor a payment card are required. Simply create an account using your email to get going.

I rapidly and easily explore with ML using Amazon SageMaker Studio Lab. You need a tool to explore with ML, whether you’re a student taking ML courses or a seasoned data scientist running several TensorFlow models.

(Note: I am not a data scientist or ML developer by trade; I am a product manager on the Amazon SageMaker team. Like you, I am always learning, breathing, and building all things ML. Nevertheless, you ought to treat my comments with caution.)

Why ought I to utilise it?

It is straightforward and tranquil. SageMaker Studio Lab enables you to devote more time to mastering machine learning (ML) rather than building the foundation for data science. Continue reading if you want to see how this differs from utilising other free notebook options or creating your own Jupyter notebook on your laptop.

Creating and maintaining an environment from scratch is difficult. Setting up a data science environment is not something you want to do for too long.

To explore with ML, you don’t want to sign up for a cloud provider and spend money on computing.

You want a setting that remembers your progress even if you log out of your account and avoids making you start over each time you log in.

To fast prototype, you want to link from the environment to services like GitHub and Kaggle.

Image from Peter Conlan’s Unsplash page

Let’s try it out.

You must first submit a new account request on the SageMaker Studio Lab website. Once your account is authorised, you can sign in after being put on a queue.

The Project interface, which includes buttons to launch the runtime, open projects, and execute samples, will appear when you log in. Decide on Open Project. Once there, you may create notebooks, terminals, source code, Markdown files, and much more.

To examine the setup, let’s open a Terminal (File > New > Terminal or select the Terminal icon).

Let’s examine the CPU data. Run the following command if your runtime’s compute type is set to CPU.

By switching the compute type on the project landing page from CPU to GPU (or vice versa), the runtime will resume using the new hardware. Once you’ve selected GPU as your computing type, return to the terminal window and issue the following command.

Let’s now verify the data in the RAM.

The most recent JupyterLab open-source package that we are using is JupyterLab 3.2.4 as of this writing. The moment has come to develop and use an ML model now that we have seen the fundamental hardware and software setup.