Python Block Mode

In the Model Training Inference Library of Python Block Mode, models trained in the AI Toolbox can be deployed on Linux hardware such as the UNIHIKER M10. ONNX models are invoked via drag-and-drop blocks, balancing ease of use with flexibility:

Module	Input	Output	Typical Use Case
Image Classification	Image / Camera	Category label + Confidence	Expression recognition, object classification
Object Detection	Image / Camera	Target position + Label + Count	Product detection, people counting
Instance Segmentation	Image / Camera	Pixel-level mask + Label	Defect detection, image segmentation
Temporal Pattern Recognition	Continuous sensor data	Pattern category label + Confidence	Motion recognition, gesture detection