ONNX Compat issues with latest skl2onnx

In order for models to work against SQL Edge the target_opset parameter needs to be set

Author: arunguru-msft

articles/azure-sql-edge/deploy-onnx.md

@@ -179,6 +179,9 @@ onnx_model_path = 'boston1.model.onnx'
 onnxmltools.utils.save_model(onnx_model, onnx_model_path)
 ```
 
+> [!NOTE]
+> You may need to set the target_opset parameter for skl2onnx.convert_sklearn function, if there is a mismatch between ONNX runtime version in SQL Edge and skl2onnx packge. See [SQL Edge Release notes](https://docs.microsoft.com/en-us/azure/azure-sql-edge/release-notes) to get the ONNX runtime version corresponding for the release and pick the target_opset for ONNX runtime based on the [ONNX backward compatibility matrix](https://github.com/microsoft/onnxruntime/blob/master/docs/Versioning.md#version-matrix).
+
 ## Test the ONNX model
 
 After converting the model to ONNX format, score the model to show little to no degradation in performance.