Machine learning models supported

The table below lists all supported model types, including the required license and software version.

Selecting a model

An introduction explaining which criteria you should consider when selecting a model can be found here: Machine Learning Cheat Sheet: selection of models.

ONNX export for supported model types

Python examples of the ONNX export from different frameworks are given for all supported model types in the section Samples of ONNX export.

Required license for supported model types

The required TwinCAT license differs depending on the model type that is loaded into the Machine Learning Runtime. Note that the TF3810 license contains the TF3800 license, which means that if the TF3810 license is valid, all models that require a TF3800 or TF3810 license can be loaded.

Supported models

For licensing, refer also to: System requirements.