Skip to main content


APIs for performing inference with Max Engine.


  • engine: Information about Max Engine, such as version.
  • model: An abstracted instance of model from any framework loaded with Max Engine InferenceSession and is ready for execution.
  • session: The entry point to MAX Engine and can be used to load models for inference.
  • tensor: Implements functionalities useful for giving input to Max Engine model for execution.
  • tensor_map: Holds the input and output tensors for a model.
  • tensor_spec: Holds a description of the input/output tensor, given or produced by Max Engine. This is similar to TensorSpec in the Mojo standard library, but is specific to Max Engine.
  • value: Types for values that can be passed in and out of engine models.