Bring Your Own Model (k-means) - minor issue

[samsammurphy]

This is a minor issue, when read the following cell..

Convert

The model format that Amazon SageMaker's k-means container expects is an MXNet NDArray

It might be helpful to the reader if the paragraph starts with a quick explanation, e.g. something like..

Convert

We created a container in advance and uploaded it to ECR using Sagemaker. You can upload a wide variety of containers. This one is setup to run the scikit learn k-means clustering algorithm and expects an MXNET NDArray ..

That is assuming I understood correctly (?)

Cheers

[djarpin]

Hi @samsammurphy - "This one is setup to run the scikit learn k-means clustering algorithm" is not actually the case. The container this example points to is the Amazon SageMaker k-means container, which uses a different methodology than scikit k-means. The purpose of this notebook was to show how you could piggyback on the inference portion of a container that already exists. Hope that clears things up, but let me know if you have any questions. Thanks.

[samsammurphy]

I see. Thanks!