Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Introduction to Model Registry

Introduction to Model Registry

「モデルのバージョン管理どうするのがいいですか」と聞かれることが多くなってきたので作成した資料です。

Asei Sugiyama

October 13, 2022
Tweet

More Decks by Asei Sugiyama

Other Decks in Technology

Transcript

  1. Introduction to Model Registry

  2. None
  3. None
  4. None
  5. None
  6. None
  7. None
  8. None
  9. None
  10. None
  11. None
  12. None
  13. import json import model_card_toolkit as mctlib model_card_json = { 'model_details':

    {'name': 'Census Income Classifier'}, 'model_details': {'overview': 'This is a wide and deep Keras model...'}, 'model_details': { 'owners':[{"name": "Model Cards Team", "contact": "[email protected]"}]}, 'considerations':{ 'use_cases':[{ "description":'This dataset that this model was trained on...'}] }, 'considerations': { 'limitations': [{ 'description':'This is a class-imbalanced dataset...'}] }, 'considerations': { 'ethical_considerations': [{ 'name': 'We risk expressing the viewpoint that...', 'mitigation_strategy': 'As mentioned, some interventions may need to be ...' }] } }
  14. None
  15. None
  16. None
  17. None
  18. None
  19. None
  20. None
  21. # Specify the model source model_url = "s3://your-bucket-name/model.tar.gz" modelpackage_inference_specification =

    { "InferenceSpecification": { "Containers": [ { "Image": '257758044811.dkr.ecr.us-east-2.amazonaws.com/sagemaker-xgboost:1.2-1', "ModelDataUrl": model_url } ], "SupportedContentTypes": [ "text/csv" ], "SupportedResponseMIMETypes": [ "text/csv" ], } }
  22. create_model_package_input_dict = { "ModelPackageGroupName" : model_package_group_name, "ModelPackageDescription" : "Model to

    detect 3 different types of irises (Setosa, Versicolour, and Virginica)", "ModelApprovalStatus" : "PendingManualApproval" } create_model_package_input_dict.update(modelpackage_inference_specification)
  23. None
  24. None
  25. None
  26. None
  27. None
  28. None
  29. None
  30. None