aws / aws/amazon-sagemaker-examples
New version of Joblib not available
- Dominant language
- Jupyter Notebook
- Stars
- 11k
- Forks
- 7k
- Avg merge
- 8h 29m
- Merged PRs (30d)
- 8
Description
My production environment uses scikit-learn(0.21.3), where I load my models with the latest PyPI joblib module. However Sagemaker supports scikit-learn(0.20.0) and uses the joblib module in sklearn.externals to dump the model which is now deprecated as per scikit-learn(0.21.3), resulting in an error in loading the model in production environment. Is there a way in which I can dump models in a sagemaker training job using PyPI's joblib instead of sklearn's joblib to ensure compatibility?
Contributor guide
Research direction
Start with the SageMaker training job and the scikit-learn 0.20.0 and 0.21.3 environments described in the report. Reproduce the model dump and load path using the mentioned joblib variants, then document or demonstrate a compatible approach for loading the resulting model in production.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- aws, scikit-learn
- Domain
- cloud, machine-learning
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100