I am following the guide to setting up a sage maker instance with the provided cloudformation template, and it seems that I am getting access denied to the S3 bucket that is housing the yaml file. Did I miss something? Was I supposed to request the permission somewhere?
Thanks for any help!
Hi James - Something isn’t right in the CloudFormation script as it fails to load. If your intention is to get a Fastai environment on Sagemaker the setup process is easy via the console. Create a Sagemaker notebook with platform identifier type of al2-v1 and instance type like g4d.xlarge (family) like in this picture below. You need to have a role with permissions or create one. VPC deployment or alternate.
Once the instance boots up use the conda_pytorch_p38 environment that has fastai1 installed. In your noteboook you can !pip update fastai or point to fastbook to get the updated fastai.
Also, how would I go about getting the course notebooks added to the instance?
In Sagemaker when you open Jupyter Labs environment, it has a GUI client to git. You can use git clone via GUI or terminal window all inside the Labs environment. Data is copied to your S3. Ensure you have adequate space. Many of the download images part requires 100GB space to download.