0

I know, one can have multiple models on a single endpoint(Multi-Model endpoints). But, can I have these multiple models of different use cases and different datasets from the s3 bucket? because we mention model artifacts while invoking for prediction. can I have?

sachin yadav
  • 1
  • 1

1 Answers1

2

Yes. You can use unrelated models. The S3 path you pass in the prediction request is the location the model artefact. Check the doc for details:) https://docs.aws.amazon.com/sagemaker/latest/dg/multi-model-endpoints.html for details.

Julien Simon
  • 2,605
  • 10
  • 20
  • 1
    The multiple models that you plan to tie together to a single endpoint should all belong to the same framework. You cant have one model belonging to tensorflow and another one to pytorch. In case ur requirement is that, you need to explore multi container endpoints on sagemaker – wingsforever Oct 27 '21 at 18:14