Multi-model Cascading Inference

The basic procedure of multi-model inference is similar to that of single-model inference. For details, see Model Management.

The differences are as follows:

• The model loading API needs to be called separately for each model in multi-model scenarios. For details about how to load a model, see Model Loading.
• If you need to run more models, call the model execution APIs separately for each model. For details about how to run a model, see Model Execution.

  For example, call the acl.mdl.execute API to implement synchronous model inference.