Can't afford to use GPU for graduation? Amazon Cloud SageMaker is free for you to use

Over the past two years, three groups of people have been particularly sensitive to graphics card prices.

One is the game party, the second is the mining party, and the other is the alchemy party. The so-called "alchemy" is a big man engaged in academic research, or a computer major.

In extreme cases, if you can't get a GPU to do machine learning, it might affect your graduation.

I have to say that graphics cards are really lacking. Not long ago, a Weibo post by Professor Chen Yiran from the Department of Electronic and Computer Engineering at Duke University also mentioned graphics cards.

He said a friend wanted to request some free GPU resources from AWS. He even laughed at himself, "People in academia not only have a lot of things to do, but they can't provide people with any profit." Graphics cards are already lacking to this point.

And I, as a senior wool party of cloud resources, used AWS Freetier resources for two years as a new user, but AWS Freetier does not have GPU resources, and it is not difficult to understand, after all, GPUs are too expensive.

A few days ago, another friend asked me, she has a friend who wants to do a complete design and needs to use a GPU, and wants to know if AWS has similar resources. I first said that there is really no such thing.

Then, I remembered that there was an AWS expert in my circle of friends, and the expert also said nicely, "There really is!"
The link is here:

• https://studiolab.sagemaker.aws/

When I saw it, this thing was completely free.

The enthusiasm of the wool party in my blood was rekindled, and I couldn't help but try it.

All kinds of basic information, the mailbox is QQ mailbox, and no credit card information is needed, hahaha!

submit application!

After verifying the email, I waited for the mysterious staff of AWS to pass my application.

I applied on Thursday at 6:00 pm and received a notification that my application was approved at 7:42 am on Friday.

【Smile】

Click to create an account, then verify your email again, and you're done!

You can choose CPU or GPU resources. In order to distinguish it from paid resources, the CPU will interrupt the runtime and process every 12 hours, but the progress will be saved, and the GPU will be interrupted once every 4 hours.

Click to start the runtime, wait for a while, the button on the right lights up.

Then there is the familiar Jupyter Notebook!

SageMaker Studio Lab is a free notebook development environment that provides 15GB persistent storage. When using, all notebooks, source code, files, and datasets can be automatically saved, and the original progress can be continued every time it is reopened.

It itself is based on the open source JupyterLab, so, how to use JupyterLab locally, you can use SageMaker Studio Lab here.

I just opened the CPU version. I tried the GPU version, and a prompt popped up. After confirming the use of the GPU, it prompted that there are currently no GPU resources available, and it may be available after waiting.

If it's a student party with little money, why don't you wait?