NEWS

---

sagemaker.mlcore 0.3.0

Architecture Change:

• moved all Amazon machine learning framework classes to sagemaker.mlframework
• moved clarify classes and method to sagemaker.common ** NOTE:** this changes reduce package overall size from 5Mb to 3.7Mb align to CRAN package size limitations.

sagemaker.mlcore 0.2.4

Bug Fix:

• Class HyperparameterTuner correctly validates hyper parameters
• Correctly pass static hyperparameters

Feature:

• Class HyperparameterTuner create method is now callable at R6ClassGenerator level. To help mimic python equivalent.

sagemaker.mlcore 0.2.3

Bug Fix:

• correctly parse parameters class

Minor:

• Add .module field to estimator classes to minic python cls.__module__

sagemaker.mlcore 0.2.2

Bug Fix:

• allow container_log_level to pass any string instead of restricting to logging levels.

Minor:

• explicitly reference sagemaker.core functions to help with mock functionality

sagemaker.mlcore 0.2.1

Bug Fix:

• Ensure deserializer classes correctly parse raw vectors as expected.

sagemaker.mlcore 0.2.0

Feature:

• Add SparseMatrixSerializer serialize class
• Allow all serializer classes read from data in from file.
• LibSVMSerializer class now uses readsparse package in the backend

Bug Fix:

• CSVSerializer correctly serialize data

sagemaker.mlcore 0.1.0

Initial release to r-universe