Get connected to a GPU deep learning server (Lesson 1)

In Part 1, Lesson 1, I’m looking under: Getting a GPU Deep Learning Server on this page:

Blockquote

The best choice of GPU servers to use with this book will change over time, as companies come and go and prices change. We maintain a list of our recommended options on the book’s website, so go there now and follow the instructions to get connected to a GPU deep learning server.

Blockquote

When I click the link to the book’s website, I’m back where I started. I don’t see a list of recommended options. Does anyone know what this link is referring to?

Thanks in advance!

Hi Bonnie

If you are just starting use Colab and pick any of the GPU not the CPU or TPU. There are three options for Colab free, cheap and expensive. If you choose to pay that is the time to consider other providers.
Colab has no permanent storage but you can put your data on your Google Drive and access it from Colab.

Regards Conwyn

Thanks for the response! I was just trying to follow the link in the lesson. For the first exercise in Lesson 1, I’m in Kaggle. When I run the first cell, this appears at the top.

image

And this at the bottom:

image

I don’t recall seeing a place to choose it.

Hi Bonnie
Sorry I thought you were using Colab from your first screenshot.
Regards Conwyn

If you click the three dots on the right of the “Draft Session” icons you can go to “Accelerator” and then pick a GPU/TPU. I always use “GPU T4 x 2”:

The course website currently does not have the list of GPU cloud options, but an old course version’s site has options listed under “Notebook Servers” and “Linux Servers”. Jeremy also walks through some options in the Lesson 0 video from a previous version of the course.

I have been able to do almost everything in the course so far (I just finished Lesson 6) in either Kaggle or Google Colab. Sometimes I run out of compute units in Google Colab, or run out of my 30 hours quota in Kaggle, so I use Paperspace which has a “Pro” subscription for $8/month which gives you access to 16GB GPUs that you can use for 6 hours at a time.