Traditional machine learning models using algorithms such as decision trees, random forests, and support vector machines, among others, can be converted to ONNX format. Such models may be produced in other environments and deployed through Oracle Database.

Once such models are converted to ONNX format, they can be deployed directly in Oracle Database and use the ONNX Runtime for inference through the SQL prediction operators. These models are typically used for tasks such as Classification, Regression, and Clustering.