Sagemaker

SageMaker is a service in AWS that can accelerate data analysis of the enormous volume of data freely available in the pacific-sound registry using machine learning. SageMaker allows for collaborative sharing of jupyter notebooks, easy ways to scale compute and storage resources for the problem at hand, and a seamless path to move from development to production.

To use SageMaker:

  1. Setup an Amazon cloud account if you don't already have one. You will only be charged for the service you use. Here is a pricing calculator.
  2. If you are new to SageMaker, read this.
  3. Follow these instructions to launch a SageMaker Studio session.
  4. Once you have your session open, this repository can be retrieved as follows:

Select the repository icon, then select "Clone a Repository"

studio_git_entry

Enter https://github.com/mbari-org/pacific-sound-notebooks.git

studio_clone_git

You should see the repository folder pacific-sound after the notebooks are checked out

studio_pacificsound_folder

The recommended kernel to start with is the Data Science kernel. You can also change the kernel at any time in the notebook.

kernel-tip

When you are done, it is recommended that you shut down everything, unless you have something that you need to run longer to avoid costs.

studio_shutdownall

Important You will not lose your data until you remove the studio session altogether. It does not delete itself after shutting down. The session can be launched again easily with your current code and data. There is a cost associated with keeping the notebooks available for the elastic storage that stores the studio environments. The cost is nominal for storing the code; be cautious about how much data you store in your studio, as elastic storage can be expensive for large collections.