The inference engine provides the following capabilities:
- TinyML optimization — Supports deployment of ultra-small models on edge devices with limited computing power, such as microcontrollers and IoT devices.
- Efficiency — Enables fast execution with low power consumption, suitable for real-time and resource-constrained applications.
- Cross-platform support — Works with different hardware platforms (for example, Cortex-M0, Cortex-M4, Cortex-M33, and Axon NPU) depending on the target device.