Setting up and accessing SageMaker API

Technical Topics PyScript
1

I have created a sagemaker env and installed all the necessary libraries - pandas, boto3, sagemaker. I can to Environment, select my sagemaker env and open iPython and then import all these three libraries with no errors. But when I “open with Jupiter notebook” and select the “sagemaker” kernel and then type and execute “Import pandas as pd” in a cell I get this error
"ModuleNotFoundError Traceback (most recent call last)
Input In [1], in <cell line: 1>()
----> 1 import pandas as pd
2 import boto3

ModuleNotFoundError: No module named ‘pandas’

Get similar errors for boto3 and sagemaker yet in my Anaconda Navigator environment, I see boto3 and sagemaker under installed column.
I also make sure awscli 2.0 is installed.

I thought those pandas, boto3 and sagemaker libraries would be available when I open a Juypter notebook and start sagemaker kernel? What am I missing here? How do I fix this issue? Thanks.
– vilas

2

dear Vilas_Tulachan1.

thank you for your posting.
This problem occurs when using multiple services simultaneously.
In each service (AWS for SageMaker, Google for Colab), services using similar libraries differ in package selection and linking, library configuration, etc.

To be precise, AWS can provide exactly the same service as anaconda, and google can also provide it. However, their commands, packages and library management standards are different.

Therefore, if installed in a different environment, it may not work due to version mismatch, repository mismatch, etc.

Regards, you.
ktsh.tanaka.2020