ann from the Spell team We’d excited to introduce our cloud GPU offering for fast.ai students.
We spent some time getting the fast.ai notebooks to work easily with Spell. We’re fast.ai learners ourselves and wanted to make sure to make it super simple for learners to get up and running with the fast.ai notebooks.
We’re really proud to offer what we think is a simple, fast, and reliable cloud GPU service that’ll keep up with your machine learning needs.
To run the notebooks, sign up for a Spell account and install the CLI tool in terminal. You’ll automatically activate your $100 credit once you sign up and enter your billing information. We email you when you have $20 credit left (and you can set additional custom notifications) so you’ll never be hit with unexpected charges.
To get started, just run the following commands in terminal (more detail in our docs):
Jupyter will be running locally, but be wired to a kernel with a K80 and the fast.ai conda environment loaded. Once Jupyter starts, switch to the “Spell - fastai on K80” kernel by clicking “Change kernel” under the Kernel menu (you only have to do this once).
Remember to stop your session (^C ^C in terminal) so you don’t get charged for an open notebook you’re not actively using!
Spell is a command line tool for super simple remote execution - like the bash & operator, but for sending work off to a cloud instance. It’s great for developing for deep learning locally but seamlessly running on a cloud GPU.
Questions/feedback welcome - we’ll be hanging out here and in slack!
Yup! We’re approving everyone off the waitlist as quickly as we can while we put the final touches on open sign ups. The waitlist should be completely removed in the next day or two.
I just approved you so you should be able to access your account now. Let me know if you have any questions getting started!
We’re really looking forward to hearing your feedback
Thanks @s_ann or letting me know. I’ve PM’ed you my feedback.
Just to double check, the $100 sign up credit is only for CPU as it says on the billing page and not for GPU as it says on this forum heading, correct?
Hi @zszazi, we changed the credits once we launched our new platform in January of this year. We now offer $10 GPU credit on sign up and $15 for you + a friend for every friend you refer. CPU use is still free for Community users.
We’ve added a bunch of new features including Jupyter Notebooks+Jupyter Lab for Community users, and features for Teams users such as Hyperparameter Search, Model Serving, Distributed Training, and Spot Instances to name a few.
We’d love for you to check out the platform and tell us what you think!