Accelerated Data Science 2.6.3 is released
- Services: Data Science
- Release Date: August 05, 2022
The following changes were made in this version.
- Added
prepare_save_deploy()method to theGenericModelclass. Now you can prepare model artifacts and deploy the model within one command. - Added support for binary payloads in model deployment.
- Updated
AutoMLModel,GenericModel,LightgbmModel,PyTorchModel,SklearnModel,TensorflowModel, andXgboostModelclasses to support binary payloads in model deployment. - To limit job runtime, added the
with_maximum_runtime_in_minutes()method in theCondaRuntime,DataFlowNotebookRuntime,DataFlowRuntime,GitPythonRuntime,NotebookRuntime, andScriptRuntimeclasses. - Deprecated the
ads.dataflow.DataFlowclass. Use theads.jobs.DataFlowclass instead. - The
ads.jobs.DataFlowclass supports published conda environments.
For more information, see Data Science, ADS SDK, and ocifs SDK. Take a look at our Data Science blog.