ImportError: cannot import name 'get_unsupported_framework_version_error'

user_6645597 0
Viewed 768

I am trying to learn AWS Machine Learning from this AWS tutorial

I have not changed any code.

When I execute the line below

from sagemaker.sklearn.processing import SKLearnProcessor

I get error:

ImportError                               Traceback (most recent call last)
<ipython-input-13-dd0123cd2e7d> in <module>
----> 1 from sagemaker.sklearn.processing import SKLearnProcessor
      2 
      3 sklearn_processor = SKLearnProcessor(framework_version='0.20.0',
      4                                      role=role,
      5                                      instance_type='ml.c4.xlarge',

~/anaconda3/envs/python3/lib/python3.6/site-packages/sagemaker/sklearn/__init__.py in <module>
     14 from __future__ import absolute_import
     15 
---> 16 from sagemaker.sklearn.estimator import SKLearn  # noqa: F401
     17 from sagemaker.sklearn.model import SKLearnModel, SKLearnPredictor  # noqa: F401
     18 from sagemaker.sklearn.processing import SKLearnProcessor  # noqa: F401

~/anaconda3/envs/python3/lib/python3.6/site-packages/sagemaker/sklearn/estimator.py in <module>
     18 from sagemaker.estimator import Framework
     19 from sagemaker.fw_registry import default_framework_uri
---> 20 from sagemaker.fw_utils import (
     21     framework_name_from_image,
     22     get_unsupported_framework_version_error,

ImportError: cannot import name 'get_unsupported_framework_version_error'
amazon-sagemaker scikit-learn amazon-web-services
1 Answers

I had the same problem. The issue with the tutorial itself and this part at the beginning:

if int(sagemaker.__version__.split('.')[0]) == 2:
    !{sys.executable} -m pip install sagemaker==1.72.0
    print("Installing previous SageMaker Version. Please restart the kernel")
else:
    print("Version is good")

It uses the old version of sagemaker library, at the same time this commit removed 'get_unsupported_framework_version_error' https://github.com/aws/sagemaker-python-sdk/commit/e4485b77991041b214e4860f3999addfee8898fa

just upgrade sagemaker to a newer version, e.g: 2.6.0

user_16652740
Related