TwinCAT Machine Learning Model Manager
The TwinCAT Machine Learning Model Manager is used to manage the ONNX model files and prepare them for use with TwinCAT. The functions of the TwinCAT Machine Learning Model Manager are summarized:
- Creation of a metadata file (JSON file).
- Creation of a PlcOpenXml, which describes a PLC data type description of the model input and output.
- Optional: Inserting application-specific metadata (Custom attributes). A user can add the Custom attributes to the JSON. The attributes can then be read out in the PLC at runtime using methods at FB_MlSvrPrediction.
The TwinCAT Machine Learning Model Manager has three different interfaces:
- 1. A graphical user interface (Visual Studio plugin)
- 2. A Python interface (Python package)
- 3. A Command Line Interface (CLI)
These are described in more detail below.
Further Information