what is the difference between using amazon sagemaker python SDK vs AWS SDK for python (Boto3)

Question

I am new to AWS, and just started to learn sagemaker, and I found there are two ways to train a model: train a model using sagemaker

1. Create and Run a Training Job (Amazon SageMaker Python SDK)
2. Create and Run a Training Job (AWS SDK for Python (Boto 3))

Is there anyone can explain what is the difference between this two ways and is there advantage and disadvantage of both ways?

Thanks

Answer 1

the SageMaker SDK is a simple, high level SDK focused on ML experimentation. It's completely abstracting infrastructure complexity, and is definitely the one to use in notebooks.

Boto3 (and other language SDKs) are service-level SDKs, giving you full control over 100% of the SageMaker API, just like you would get on EC2, etc. This comes at the cost of lower level operations, more verbosity, etc. This SDK is best for production, automation, etc.

Hope this helps.