example-get-started-experiments: redeploy to same endpoint

[dberenbaum]

The model deployment action breaks when we redeploy the repo because the endpoint already exists with the same name from the old repo, and it looks like aws has no easy solution for updating an existing endpoint. Not critical since we don't use the endpoint now, but it makes the model deploy action look broken.

[daavoo]

It is possible to delete the existing endpoint and configuration and then the same workflow would succeed.
We could do the deletion as part of the github workflow or the generate.sh script:
https://docs.aws.amazon.com/sagemaker/latest/dg/realtime-endpoints-delete-resources.html

[dberenbaum]

Seems this still isn't working: https://github.com/iterative/example-get-started-experiments/actions/runs/6461193502/job/17540424484