Google Cloud Platform $300 free credit — incredible offer for Deep Learning students

In case you have not already signed up for GCP account, you can enable GCP from any of your gmail account from here and get $300 free credit for first year. This credit is good enough for almost 1000 hours of Nvidia Tesla K80 GPUs!!

You will need to setup your billing info here but Google won’t charge you anything unless you run out of your free credits and give your explicit consent for charge.

Clouderizer can link with your GCP account to allow you run your deep learning projects (like course project) seamlessly from within Clouderizer console itself. Once you have an signed up for GCP, login to your Clouderizer console and press Start for any of your existing project. In case you don’t have any project, you can create a new project and then press Start.

Select Google Cloud Platform from the list. It should prompt you to setup GCP integration.

Follow the link and press Authorize Access to Google Cloud Platform

This should take you to Google OAuth workflow. Sign in with same gmail id you used for signing up for GCP above. You might see a warning from Google saying Clouderizer is not verified yet.

We are currently in review process with Google and it might take few weeks before this warning can be lifted. Meanwhile, you can click on link Go to Clouderizer (unsafe) to proceed further. It will take you to authorization screen where you can allow Clouderizer to manage GCP resources on your behalf.
Please note at any point of time, you can login to your Google account and visit this URL to revoke this access

Once you Allow, you will be taken back to Clouderizer console with GCP setup successful message.

Now again you can press Start on your project and select Google Cloud Platform. You will now be presented with various GCP machine options to run your project on. You can press Start button for the machine option of your choice.

When running for the first time, you might see an error like this

This is due to GCP Compute Engine APIs not enabled yet for Clouderizer. This error message will give you a URL to enable this API. Copy and visit this URL and follow on-screen instructions to enable this API. This might take few minutes if you have signed up for GCP just now.

Once Compute Engine APIs are enabled, you can come back to Clouderizer console and try starting the project again.
It should now trigger automated process of creating a new VM instance in GCP, setting up your project environment, downloading your code and datasets. You can track the progress of your machine setup from Clouderizer console.

Once your project setup is ready, its status should change to Running. You can now use JupyterLab and remote terminal button to work on your project.

**Important Note

Some users, who have signed up for GCP recently, can see this error on starting their projects

“Quota ‘GPUS_ALL_REGIONS’ exceeded. Limit: 0.0 globally.”

Please go to Compute Engine -> Quotas -> Iam and admin (hyperlink) and search for the following quota
GPUs (all region).

In case limit for this is 0, you need to request Google to increase this to 1 or whatever value you need.

Originally published at