Workflows for users

Workflow overview

The TwinCAT Machine Learning Creator provides a complete workflow for creating and deploying AI models within a web-based user interface.

The workflow consists of the following steps:

Create project

First, a project is created on the platform. The project defines the task types (e.g., image classification) and serves as a framework for data sets and models.

Create a data set

First, image data is uploaded to the platform. The images are then (or immediately upon upload) tagged with labels that define the desired target class. The data set forms the basis for subsequent model training.

Configure model training

For training, a new model is created and linked to a data set. In addition, training-related parameters, as well as target hardware, target software, and runtime requirements, can be defined.

Run a model training session

Once training starts, the platform automatically generates an AI model based on the training data provided.

Validate and analyze the model

Once training is complete, various analysis and explainability functions are available. These include, among other things, statistical evaluation methods, confidence intervals, confusion matrices, and visual representations that provide insight into the model's decision-making process.

Export model

Trained models can then be downloaded. In addition, PLCopen-XML can be exported to simplify integration into TwinCAT projects.