With the Vitis AI runtime, applications access a unified high-level runtime API that facilitates seamless and efficient data center-to-Edge deployments. This cohesive interface streamlines the application deployment process, ensuring a smooth transition between data center and Edge environments.
The following are the features of the AI runtime API:
- Asynchronous submission of jobs to the accelerator
- Asynchronous collection of jobs from the accelerator
- C++ and Python implementations
- Support for multi-threading and multi-process execution
The following figure depicts the VART framework. XIR is the AMD Intermediate Representation, an internal IR used in Vitis AI to represent neural network operators.
Figure 1. VART Stack
