Main new features:
AI SDK now allows TensorRT optimization to be configured directly when model is added to the GPU component
AI SDK allows the creation and testing of the Edge package without the need to create and convert and configuration package
GPURuntime.use_model() method saves the model with its original name and ensures the compatibility with the AI Inference Server
AI SDK will raise an error if reserved keywords are used as input variable names
Deprecated features:
The following classes has been deprecated. Use
pipeline.export()instead to create an edge configuration package directly.simaticai.deployment.Pipeline.save()simaticai.deployment.convert_packagesave()
Fixed issues:
none