I may have missed a step. I connected to the EC2 instance via ssh, and then executed jupyter notebook. I am pretty sure that Jupyter Notebook doesn’t come with Linux - you typically have to install it.
So, how did Jupyter Notebook get on my EC2 instance?
I’m assuming you installed anaconda in the process of setting up the server?
The anaconda distribution comes bundled with jupyter notebook.
I recommend using the “which” command to see where your binary is running from. As in “which jupyter”
To add to @NegatioN’s response, if you’re using the fast.ai AMI to set up the EC2 instance, Jeremy and Rachel have pre-installed Anaconda, Cuda and the Nvidia drivers on it for you.
Are you saying that Amazon actually designed one of their mass market boxes to Jeremy and Rachel’s specs?
I did but that was for my local cygwin. I didn’t install anything on the ec2 instance.
If you used the script provided by Jeremy and Rachel ( https://github.com/fastai/courses/blob/master/setup/setup_p2.sh ), you are not using a standard “untainted” installation provided by amazon. But one with anaconda already installed and put on your path. Which means it’s instantly accessible.
Please run “which jupyter” on the ec2 instance, and confirm or refute this, otherwise it’s very difficult to help you, as we have no information to go on.
Take a look at what an AMI is : https://en.m.wikipedia.org/wiki/Amazon_Machine_Image
In short, an AMI is a read-only filesystem image that includes an operating system (e.g., Linux, Unix, or Windows) and any additional software required to deliver a service or a portion of it.
Jeremy and Rachel created an AMI containing the exact environment they used - including Anaconda, Jupyter notebook and the drivers and made it available on Amazon for us. So now we can choose to deploy our EC2 instance based on this AMI and have the exact same environment on our instances.